3 and MBL/Ficolin/CL-11 Associated Serine

Home , Collectin

Heterocomplex Formation between MBL/Ficolin/CL-11−Associated Serine Protease-1 and -3 and MBL/Ficolin/CL-11− Associated Protein-1 This information is current as of October 1, 2021. Anne Rosbjerg, Lea Munthe-Fog, Peter Garred and Mikkel-Ole Skjoedt J Immunol 2014; 192:4352-4360; Prepublished online 28 March 2014;

doi: 10.4049/jimmunol.1303263 Downloaded from http://www.jimmunol.org/content/192/9/4352

References This article cites 39 articles, 22 of which you can access for free at: http://www.jimmunol.org/content/192/9/4352.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

by guest on October 1, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2014 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Heterocomplex Formation between MBL/Ficolin/ CL-11–Associated Serine Protease-1 and -3 and MBL/Ficolin/CL-11–Associated Protein-1

Anne Rosbjerg, Lea Munthe-Fog, Peter Garred, and Mikkel-Ole Skjoedt

The activity of the complement system is tightly controlled by many fluid-phase and tissue-bound regulators. Mannose-binding lectin (MBL)/ficolin/collectin-11–associated protein-1 (MAP-1) is a recently discovered plasma protein that acts as an upstream inhibitor of the lectin complement pathway (LCP). It has previously been shown that MAP-1 can compete with the MBL/ficolin/ collectin-11–associated serine proteases (MASPs) in binding to MBL and the ficolins. However, this mechanism may only partly explain the inhibitory complement effect of MAP-1. We hypothesized that MAP-1 is also involved in heterocomplex formation with the MASPs thereby breaking the stoichiometry of the activation complexes of the LCP, which could represent an alternative mechanism of MAP-1–mediated complement inhibition. We assessed the heterocomplex formation with ELISA, size-exclusion Downloaded from chromatography, and immunoblotting using both recombinant proteins and serum/plasma. We found that rMAP-1 can engage in heterocomplexes with rMASP-1 and rMASP-3 in a calcium-dependent manner. Moreover, we discovered that rMASP-1 and rMASP-3 also form heterocomplexes under these conditions. Complexes containing both MAP-1 and MASP-1 or -3 were detected innormalhumanserumandplasma,anddepletionoftheLCPrecognition molecules from ficolin-3–deficient human serum showed that free circulating heterocomplexes also exist in the blood, although the major part appears to be associated with the

LCP recognition molecules. Altogether, these ﬁndings suggest that MASPs can associate in various combinations and bring new http://www.jimmunol.org/ perspectives to the complexity of lectin pathway–driven complement activation. The Journal of Immunology, 2014, 192: 4352–4360.

he lectin complement pathway (LCP) is one of three ini- tissue damage following myocardial (3, 4), gastrointestinal (5), tiating pathways in the complement cascade and is a part renal (6, 7), and cerebral (8) I/R. MASP-2 knockout mice also T of the innate immune response. The pattern-recognition demonstrate the influence of LCP in myocardial and gastrointes- molecules (PRMs) from the LCP, mannose-binding lectin (MBL), tinal I/R tissue injuries (9). The protective function of LCP in the the ficolins (ficolin-1, -2, and -3), and collectin (CL)-11 bind to early immune response can thus quickly shift and become a direct by guest on October 1, 2021 microbial surfaces and initiate the cascade through MBL/ficolin/ source of tissue damage during situations of oxidative stress. CL-11–associated serine proteases (MASPs) (1, 2). LCP is there- Hence, novel complement regulators are becoming important in fore important in the defense against intruding microorganisms. the prospect of future medical treatment in situations of excessive Inappropriate activity, in contrast, can create detrimental damages complement activation. through recognition and destruction of self-tissue. It is believed that MBL/ficolin/CL-11–associated protein-1 (MAP-1; also known LCP potentiates tissue damage in various ischemia/reperfusion as MAp44) is a newly discovered transcriptional variant of the (I/R) conditions because oxidative stress allows MBL to bind to MASP1 gene and has shown the ability to inhibit LCP activation vascular endothelium (3). In vivo experiments in rats and mice in vitro (10) and also in vivo, where MAP-1 protects against I/R have shown that MBL is a contributing factor in the generation of injuries and inhibits thrombogenesis in mice (11). MAP-1 does not contain a serine protease domain, as it is terminated in a Laboratory of Molecular Medicine, Department of Clinical Immunology, Faculty of unique amino acid sequence located after the first CCP domain. Health and Medical Sciences, Rigshospitalet, University of Copenhagen, DK 2100 However, MAP-1 comprises the H chain domains that mediate Copenhagen, Denmark both homodimerization and MBL/ficolin binding similar to the Received for publication December 6, 2013. Accepted for publication February 19, other MASPs (12–16). In line with this, we have previously shown 2014. that MAP-1 can outcompete the MASPs for MBL and ficolin-3 This work was supported by The Danish Council for Independent Research in Med- ical Sciences, the Novo Nordisk Foundation, the Svend Andersen Research Founda- binding (11, 14), and in the current study, we questioned the ca- tion, the Capital Region of Denmark, and Rigshospitalet. pability of MAP-1 to heterodimerize with the other MASPs. This Address correspondence and reprint requests to Anne Rosbjerg, Laboratory would shed new light on the mechanism of MAP-1 inhibition, but of Molecular Medicine, Department of Clinical Immunology, Section 7631, also on the configuration of the LCP activation complex, where Rigshospitalet, Blegdamsvej 9, DK 2100 Copenhagen, Denmark. E-mail address: [email protected] the possible existence of MASP heterodimers has not been firmly established. Abbreviations used in this article: CHO, Chinese hamster ovary; CL, collectin; I/R, ischemia/reperfusion; LCP, lectin complement pathway; MAP-1, mannose-binding In general, it is debated how the initial activation mechanism in lectin/ficolin/collectin-11–associated protein-1; MASP, mannose-binding lectin/fico- the LCP works. Whether the protease domains of MASPs are cis- lin/collectin-11–associated serine protease; MBL, mannose-binding lectin; NHS, normal human serum; pAb, polyclonal Ab; PRM, pattern-recognition molecule; RPMI+ activatedwithinthe same dimer (17), trans-activated between suppl, RPMI 1640 medium supplemented with 10% FCS, 100 U/ml penicillin, 0.1 separate dimers (18), or whether there is an alternative means of mg/ml streptomycin, 2 mM L-glutamine, and 200 nM methotrexate; SEC, size- activation is still not clear. The existence of MAP-1 hetero- exclusion chromatography; TBS/Ca/Tw, TBS/2 mM CaCl /0.05% Tween 20. 2 complexes could, in any case, very well influence complement Copyright Ó 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 activation, because a protease-to-protease coactivation step would www.jimmunol.org/cgi/doi/10.4049/jimmunol.1303263 The Journal of Immunology 4353 be disrupted. The observation that heterocomplex formation be- The same procedure was done with the addition of 10 mM EDTA in the tween MAP-1 and the MASPs occurs would therefore indicate the immunoprecipitation step. existence of an alternative inhibitory mechanism of MAP-1 and Size-exclusion chromatography of in vitro heterocomplexes indicate that the LCP actually comprises MASP heterocomplexes m m as well. MAP-1 (75 g) and MASP-1 or -3 (150 g) heterocomplexes were generated as described above. 300 ml complex solution were run on a Super- dex 200 HR 10/30 column (GE Healthcare) including complexes that had Materials and Methods been preincubated with 10 mM EDTA for 2 h. In addition, rMAP-1, Primary Abs rMASP-1,andrMASP-3dilutedinTBS/5mMCaCl2 or TBS/10 mM EDTA were analyzed separately. Running buffers corresponded to the In-house produced mAbs used in the following assays were: anti–MASP-1/- applied dilution buffers. Fractions of 0.5 ml were collected and used in 3 mAb F3-46 (14), anti–MAP-1 mAb 20C4 (10), anti–MASP-3 mAb 7D8, western blotting probing with a combination of biotinylated mAbs 8B3 and and anti–MASP-1/-3/MAP-1 mAb 8B3 (19), anti–ficolin-1 mAb FCN166 20C4. (20) anti–ficolin-2 mAbs FCN216 and FCN219 (21), and anti–ficolin-3 mAb FCN334 (22). Moreover, we used the following commercial Abs: Cotransfection of CHO cells anti–MASP-1 polyclonal Ab (pAb) (C-20, sc-50839; Santa Cruz Bio- CHO cells expressing rMAP-1 and rMASP-3 as previously described were technology, Heidelberg, Germany), anti–ficolin-1 pAb (HP9039; Hycult transiently transfected with a MASP-3 or MAP-1 vector, respectively. Cells Biotech, Uden, The Netherlands), anti-MBL mAbs (HYB 131-10 and 131- were seeded in a culture plate with RPMI+suppl in triplicates of 0.5 3 104 11; Bioporto Diagnotics, Gentofte, Denmark), and finally, anti–CL-11 pAb cells/well. The next day, the culture supernatant were replaced with a mix (antiCOLEC-11, 15269-1-AP; Proteintech, Manchester, U.K.). of DMEM (Life Technologies), Lipofectamine 2000 (Life Technologies), Biotin labeling and vector. Detection of heterocomplexes in the culture supernatant was performed in ELISA assays as previously described using capturing Abs Biotin labeling was done using biotin-N-hydroxysuccinimide ester (H1759; that targeted the transiently expressed heterocomplex proteins (mAbs 7D8/ Downloaded from Sigma-Aldrich, Broendby, Denmark). 20C4 or 20C4/F3-46). Thus, 7D8 was applied to capture MASP-3 in the culture supernatant from cells with a stabile MAP-1/transient MASP-3 Recombinant proteins transfection and 20C4 was subsequently used for MAP-1 detection. Con- versely, 20C4 was used as the capturer Ab for the culture supernatant with Recombinant proteins were generated as previously described (23). In short, stabile MASP-3/transient MAP-1 expression and F3-46 was applied to MAP-1, MASP-1, and MASP-3 were expressed in CHO-DG44 cells cul- detect the MAP-1-bound MASP-3. tivated in RPMI 1640 medium (Sigma-Aldrich) supplemented with 10% FCS, 100 U/ml penicillin, 0.1 mg/ml streptomycin, 2 mM L-glutamine, and Coculture of CHO cells http://www.jimmunol.org/ 200 nM methotrexate (RPMI+suppl) or by using serum-free medium Chinese hamster ovary (CHO) CD1 (Lonza, Vallensbaek, Denmark) sup- A ratio of 1:1 CHO cells expressing rMAP-1 and rMASP-3 were seeded in 4 plemented with 200 nM methotrexate (serum-free media). Purification was triplicates reaching a total of 1 3 10 cells per well in RPMI+suppl. performed with affinity chromatography using anti–MASP-1/-3 mAb 8B3 Adhering cells were afterward washed and changed into serum-free media or anti–MAP-1 mAb 20C4 as previously described (10). plus supplement. The culture supernatants were harvested the following day and a fraction of supernatant from each well was incubated with 10 mM In vitro heterocomplexes EDTA overnight. Finally, all samples were transferred to ELISA plates m m to detect heterocomplexes (mAbs 20C4/F3-46) and heterocomplexes plus Equimolar amounts of rMAP-1 (2 g/ml), rMASP-1 (4 g/ml), and MAP-1 homodimers (mAbs 20C4/8B3). rMASP-3 (4 mg/ml) were combined in the following combinations; MAP-

1–MASP-1, MAP-1–MASP-3, and MASP-1–MASP-3. The different by guest on October 1, 2021 combinations were dialyzed against TBS/EDTA (10 mM Tris-HCL, 150 mM NaCl, and 10 mM EDTA [pH 7.4]) for 2 h at 4˚C (dissociation step) and thereafter against TBS/5 mM CaCl2 overnight at 4˚C (reassembling step). In vitro heterocomplexes measured by ELISA ELISA plates were coated with 5 mg/ml mAb 20C4. Plates were blocked for1hwithTBS/2mMCaCl2 /0.05% Tween 20 (TBS/Ca/Tw) at 20˚C before overnight incubation at 4˚C with serial dilutions of MAP-1– MASP-1 or MAP-1–MASP-3 complex mixtures. Plates were washed/ blocked in TBS/Ca/Tw, and signals were obtained with H2O2 and orthophenylenediamine (DakoCytomation, Glostrup, Denmark) after 2 h of incubation with 2 mg/ml biotinylated mAb F3-46 preceding 1 h incubation with HRP-conjugated streptavidin (GE Healthcare, Broendby, Denmark). The same assay was performed on complex mixtures after 2 or 24 h of incubation with 10 mM EDTA at 20˚C and 4˚C, respectively. TBS/Tw was applied as washing/blocking buffer and TBS/Tw/10 mM EDTA as dilution buffer.

In vitro heterocomplexes detected by Western blotting ELISA plates were coated with 5 mg/ml mAb 7D8 overnight at 4˚C. After blocking in TBS/Ca/Tw, plates were incubated overnight at 4˚C with MASP-1–MASP-3 complex mixture diluted in TBS/Ca/Tw. The next day, washing and blocking of plates was followed by an addition of 200 ml TBS to all wells. TBS from the first well was afterward replaced with 10 ml 1:1 mix of TBS and LDS sample buffer. With 10-min intervals, TBS and LDS sample buffer were transferred from one well to the next, resulting in a 10-ml eluate containing content from all 12 wells in a row. The final eluate 3 from 2 12 wells was run on a 4–12% bis-Tris polyacrylamide gel (Life FIGURE 1. Formation of in vitro heterocomplexes between MAP-1 and Technologies, Nearum, Denmark) and blotted onto a polyvinylidene B A difluoride membrane (GE Healthcare) as recommended by the manufac- MASP-1 ( ) and -3 ( ). Heterocomplex generation of equimolar amounts turer (Life Technologies). The membranes were probed with pAb C-20 of rMAP-1 and rMASP-1 or -3. The complexes were traced in a sandwich followed by HRP-conjugated donkey anti-goat pAb (A00178; Genscript, ELISA, and moreover, the three proteins—rMAP-1, rMASP-1, and -3— Piscataway, NJ). Membranes were developed using SuperSignal West were applied separately as negative controls. Results are presented as Femto Chemiluminescent Substrate (Thermo Scientific, Rockford, IL). means of triple determinations 6 SD. 4354 HETEROCOMPLEXES IN THE LECTIN COMPLEMENT PATHWAY

20C4/8B3. Additionally, SEC was performed on culture supernatant from MAP-1 and MASP-3 CHO cells cultivated separately in serum-free media, and fractions were used in ELISA assays measuring MAP-1 homodimers (mAbs 20C4/8B3) and MASP-3 homodimers (mAbs 7D8/F3-46).

Heterocomplex binding to ficolin-3 Culture supernatant from cocultivated MAP-1 and MASP-3 CHO cells was incubated with a high concentration of ficolin-3 (60 mg/ml) to circumvent binding competition from homodimers. Next, ficolin-3 was pulled out using a combination of anti–ficolin-3–specific mAb (FCN334) and Pan Mouse IgG Dynabeads (Life Technologies), and heterocomplexes were detected as previously described (mAbs 20C4/F3-46). Ficolin-3 was substituted with TBS buffer in the control sample.

FIGURE 2. Formation of in vitro heterocomplexes between MASP-1 Serum and plasma complexes in ELISA and MASP-3. Shown are 4–12% SDS-PAGE and Western blotting of an Equal volumes of dilution buffer and normal human serum (NHS), EDTA MASP-3–specific immunoprecipitation (IP) of MASP-1–MASP-3 com- plasma, citrated plasma, or heparinized plasma were added in serial plexes using an MASP-1–specific Ab for immunoblot detection. Lane 1: dilutions to mAb 20C4–precoated ELISA plates that were developed as rMASP-1 homodimer. Lane 2: rMASP-3 homodimer. Lane 3: MASP-1/-3 previously described using biotinylated mAb F3-46. heterocomplex solution. Lane 4: MASP-1/-3 heterocomplex solution with 10 mM EDTA in IP step. Lane 5: rMASP-1 IP with the MASP-3–specific Depletion of ficolin-3–deficient serum Ab. The blots are representatives for three parallel experiments. 2/2 Serum from a ficolin-3–deficient (FCN3 ) patient was depleted for Downloaded from ficolin-1, ficolin-2, MBL, and CL-11 using a combination of Ab precip- 2/2 Size-exclusion chromatography of coculture heterocomplexes itation and ligand-mediated depletion (24). FCN3 serum and TBS buffer was mixed 1:1 and incubated with mannose-agarose (M6400; A total of 300 ml coculture supernatant was used for size-exclusion Sigma-Aldrich) for 4 h at 4˚C. The supernatant was collected and mixed chromatography (SEC) in the same manner as previously described, and with N-acetyl-D-glucosamine–agarose (A2278; Sigma-Aldrich) for incu- fractions were applied in an ELISA assay measuring heterocomplexes bation overnight at 4˚C. Once more, the supernatant was collected, and (mAbs 7D8/20C4). We also measured MAP-1 homodimers by immuno- mAb FCN219 was added in two rounds using concentrations of 10 and precipitating MASP-3 using mAb 7D8 and Pan Mouse IgG Dynabeads 5 mg/ml, respectively. Ag–Ab complexes were pulled out after each round http://www.jimmunol.org/ (Life Technologies) prior to SEC and assayed the fractions using mAbs with Pan Mouse IgG Dynabeads (11041; Life Technologies). A fraction of by guest on October 1, 2021

FIGURE 3. SEC of MAP-1 in vitro heterocomplexes. Heterocomplexes were generated from equimolar amounts of rMAP-1 and rMASP-1 or -3. The heterocomplex solution was fractionated into 0.5-ml fractions on a Superdex 200 HR 10/30 column under calcium conditions (A, B) and after preincubation with EDTA (C, D). rMAP-1 (dotted line) and rMASP-1 and -3 (dashed line) were also applied with either calcium or EDTA in the running buffer. The arrows in the top panels indicate that the heterocomplex solutions (solid line) generated an extra peak representing the newly formed heterocomplexes. The Western blots in the bottom panels show the formation of heterocomplexes as MAP-1 and MASP-1 or -3 makes an overlap in fraction ∼21–26 under calcium conditions (A, B), whereas EDTA reseparates the proteins into different fractions (C, D). To get a better view of the different peaks, we used a different scale on the y-axis for each run, hence no definite absorbance values were applied. The Journal of Immunology 4355 the PRM-depleted FCN32/2 serum was furthermore depleted for MAP-1, blots, in which lane 3 was loaded with the heterocomplex solution MASP-1, and MASP-3 in the same manner using mAb 8B3. Lastly, the precipitated under calcium conditions. This resulted in an 85- presence of heterocomplexes was determined (mAbs 20C4/F3-46). Veri- kDa MASP-1 band, signifying that MASP-1 and MASP-3 were fication of depletion was carried out in sandwich ELISAs: ficolin-1, FCN166-HP9039 (50% serum); ficolin-2, FCN216-FCN219* (10% se- coassociated. Meanwhile, lane 4 contained heterocomplexes pre- rum); ficolin-3, FCN334-FCN334* (1% serum); and MBL, 131.11-131.1* cipitated in the presence of EDTA, causing the MASP-1 band to (3% serum). CL-11 depletion was tested in Western blotting using anti– vanish presumably because heterocomplexes were dissociated dur- COLEC-11 (*biotinylated). ing immunoprecipitation. Lanes 1, 2, and 5 served as controls—lane 1 containing rMASP-1 as a positive control and lanes 2 and 5 Results containing negative controls: pure rMASP-3 homodimer and pure MAP-1 in vitro heterocomplexes rMASP-1 homodimer precipitated with an anti–MASP-3–specific The level of in vitro heterocomplexes comprising MAP-1 and Ab, respectively. MASP-1 or MASP-3, respectively, was assessed after dissociation SEC of in vitro heterocomplexes of the recombinant homocomplexes via calcium chelation followed by a reassembling step of the different monomers under calcium- MAP-1wasmixedwitheitherMASP-1or-3inequimolaramounts, sufficient conditions. The formation of new heterocomplexes was and heterocomplexes were generated as previously described. The measured in a sandwich ELISA assay capturing MAP-1 and Superdex 200 HR 10/30 column separated the different protein detecting MASP-1 and -3. A dose-dependent signal was obtained complexes into 0.5-ml fractions, and a continuous measurement of for the MAP-1 complex formation with both MASP-1 and -3 as the absorbance generated the peaks seen in Fig. 3. Fig. 3A and 3B illustrated in Fig. 1 [MAP-1–MASP-3 (Fig. 1A) and MAP-1– show the diagrams generated from the MAP-1–MASP-1 and Downloaded from MASP-1 (Fig. 1B)]. Pure rMAP-1, rMASP-1, and rMASP-3 MAP-1–MASP-3 complex solutions, respectively, also including homodimers were used as negative controls. We further charac- the peaks generated by pure rMAP-1, rMASP-1, and rMASP-3 terized the complex bond by incubating heterocomplex mixtures homodimers. with 10 mM EDTA for 2 and 24 h preceding analysis. The results By evaluating the different peaks in each of the two top dia- showed a decreased signal upon EDTA treatment after 2 h and an grams, it appears that an extra peak is formed in the complex almost complete loss of signal after 24 h, meaning that calcium solution compared with the peaks generated by the separate pro- chelation mediated a slow disruption of the newly formed heter- teins. The first peak in the complex diagrams corresponds to the http://www.jimmunol.org/ ocomplexes (Fig. 1). MASP-1 or MASP-3 peaks, respectively, and the third peak then corresponds to the MAP-1 peak. These two peaks flank a middle MASP-1–MASP-3 in vitro heterocomplexes peak representing a complex that sizewise corresponds to a hybrid Heterocomplex formation between MASP-1 and MASP-3 was of MAP-1 and MASP-1 or -3 (see arrows in Fig. 3A and 3B). Es- shown in Western blotting. We subjected the heterocomplex so- pecially the MAP-1–MASP-3 peak is visible. The MAP-1–MASP-1 lution to immunoprecipitation targeting MASP-3 followed by peak is more masked, as the size of this complex falls in closer analysis of the precipitates in SDS-PAGE/Western blotting with proximity of the MAP-1 homodimer, which probably exceeds the specific detection of MASP-1. This enabled us to detect whether resolution capacity of the column. In comparison, we used the by guest on October 1, 2021 MASP-1–MASP-3 complexes had been formed. Fig. 2 depicts the same setup for heterocomplexes preincubated with 10 mM EDTA for 2 h, which caused the middle peak to vanish, suggesting that the heterocomplexes were dissociated. The homodimers were also dissociated into monomers in the presence of EDTA, and this calcium-dependent dimerization of MAP-1 and the MASPs is in agreement with previous notations (14, 16, 17, 19). The high molecular structures found in the area around fractions 14–19 were not MASP derived complexes or aggregates, because

FIGURE 5. MAP-1 and MASP-3 heterocomplex formation via cocultivated CHO cells. The stably transfected CHO cells expressing either FIGURE 4. MAP-1 and MASP-3 heterocomplex formation via cotrans- rMAP-1 or rMASP-3 were mixed in cocultures and grown in serum-free fected CHO cells. Stably transfected CHO cells were transiently cotrans- media. Heterocomplexes and heterocomplexes plus MAP-1 homodimers in fected to enable coexpression of rMAP-1 and rMASP-3 from the same cell. the culture supernatant were measured in ELISA, which showed that the (A) Stable MAP-1/transient MASP-3 transfection. (B) Stable MASP-3/ amounts of heterocomplexes and MAP-1 homodimers were approximately transient MAP-1 transfection. Heterocomplexes in the culture supernatant the same (determining an exact ratio is difﬁcult). Preincubating the culture were measured in ELISA using capturing Abs that targeted the transiently supernatant with EDTA completely dissociated the heterocomplexes. expressed protein. Results are presented as means of triple determinations Results are presented as means of triple determinations (three different (three different culture wells) 6 SD. culture wells) 6 SD. 4356 HETEROCOMPLEXES IN THE LECTIN COMPLEMENT PATHWAY

FIGURE 6. ELISA proﬁle of SEC fractions from cocultivated and separately cultivated CHO cells. Culture supernatant from cocultivated MAP-1 and MASP-3 CHO was applied in SEC, and the heterocomplex and MAP-1 homodimer content in each fraction was thereafter measured in ELISA. For comparison, SEC was also performed on culture supernatants from separately cultivated MAP-1 and MASP-3 CHO cells followed by MAP-1 homodimer and MASP-3 homodimer detection in ELISA. The heterocomplex top fraction (fraction 22) was posi- tioned right in between the MASP-3 homodimer peak (fraction 20) and MAP-1 homodimer peak (fraction 25), suggesting that the MAP-1–MASP-3 heterocomplexes are dimers. the fractions were blank in Western blots detecting MAP-1, natants (Fig. 6). We were also able to detect MAP-1 homodimers MASP-1, and MASP-3. On the contrary, Western blot analyses in the coculture SEC fractions equivalent to the single MAP-1 culture showed heterocomplex formation in the peak fractions; MAP-1 (fractions 24–26). With the Abs used (20C4/8B3), it was possible to

and MASP-1 or -3 divide in separate fractions due to size dif- discriminate between the MAP-1 homodimers and the hetero- Downloaded from ference. However, if heterocomplexes between these proteins were complexes from the coculture via an initial MASP-3 depletion. formed, it would expectedly place MAP-1 and MASP-1 or -3 in The coculture SEC fractions were nondiluted in the hetero- common fractions, and we indeed observed an overlap of MAP-1 complex measurement, whereas the MAP-1 homodimers were and MASP-1 or -3 in fractions ∼21–26 (Fig. 3A, 3B). Further- measured in an 83 dilution, and the separate culture fractions more, we could see that EDTA mediated a reseparation of the were 163 diluted. But because we used different cell cultures and

proteins into different fractions (Fig. 3C, 3D). different Abs, this does not give an indication of the relative http://www.jimmunol.org/ Another EDTA-mediated effect became visible in the MAP- amounts of various complexes. The objective was to visualize the 1–MASP-1 fractions, in which EDTA caused a shift from pre- top fraction positioning of heterocomplexes versus homodimers to dominantly full-length MASP-1 bands into a-chain bands. Thus, get an indication of the protein complex composition. Seeing that H and L chain (a and b chain) were cleaved apart, which means MAP-1–MASP-3 heterocomplexes are located in between MAP-1 that MASP-1 went from a nonactivated to a cleaved/activated and MASP-3 homodimers, it seems likely that MAP-1 and state. MASP-3 have dimerized. Heterocomplex binding to ﬁcolin-3 Intracellular heterocomplex formation

We tested whether heterocomplexes in the coculture supernatant by guest on October 1, 2021 The study of MASP heteromer formation in the secretory pathway were able to bind to recombinant ficolin-3 in solution by precip- was approached by cotransfecting CHO cells, thereby enabling itating ficolin-3–bound heterocomplexes with an anti–ficolin-3 them to simultaneously express rMAP-1 and rMASP-3. This was mAb. Fig. 7 shows the level of heterocomplexes present in the carried out in two setups: stably MAP-1 transfected CHO cells were supernatant before and after ficolin-3 incubation/precipitation. transiently transfected with a MASP-3 vector and vice versa. Fig. 4 Ficolin-3 incubation/precipitation resulted in a vast reduction in shows the results from both cotransfections, in which especially the level of heterocomplexes, whereas the control sample without the combination of stable MAP-1/transient MASP-3 transfection ficolin-3 only had a small reduction compared with the non- promoted heterocomplex formation (Fig. 4A). The opposite situ- precipitated sample. Hence, the majority of heterocomplexes in ation generated new heterocomplexes as well, but in lower the culture supernatant can associate with ficolin-3. amounts (Fig. 4B). In vivo MAP-1 complexes in human serum and plasma Extracellular heterocomplex formation Because heterocomplexes were able to form from both purified The possibility of extracellular heterocomplex formation was in- recombinant proteins and in cell cultures, we assessed whether vestigated by combining MAP-1 and MASP-3 CHO cells in serum- free media, thus preventing serum-related PRMs from intervening as a linker between MAP-1/MASP-3. The results show that MAP-1 and MASP-3 form heterocomplexes in the extracellular environment. In Fig. 5, we looked at both MAP-1–MASP-3 heterocomplexes and heterocomplexes plus MAP-1 homodimers. The exact ratio between homodimers and heterocomplexes is difficult to assess with two different detection Abs. However, an Ab- affinity test showed that 8B3 had a higher affinity for MASP-3 than F3-46, (data not shown), suggesting that the level of heterocomplexes compared with homodimers might be underestimated. FIGURE 7. Heterocomplex binding to ficolin-3. Coculture supernatant Moreover, Fig. 5 illustrates that the heterocomplexes could dis- containing MAP-1–MASP-3 heterocomplexes underwent either ficolin-3 sociate by chelating calcium with EDTA. or TBS buffer incubation prior to ficolin-3 immunoprecipitation. The level AnELISAassaymeasuringonlyheterocomplexes in the coculture of heterocomplexes was drastically decreased in the ficolin-3 sample supernatant SEC fractions confirmed that a peak is generated and, compared with the TBS sample, meaning that heterocomplexes bound to moreover, located exactly in between the MAP-1 and MASP-3 and coprecipitated with ficolin-3. Results are presented as means of triple homodimer peaks originating from their separate culture super- determinations 6 SD. The Journal of Immunology 4357 these heterocomplexes have an actual physiological existence. First, we looked for complexes in general comprising both MAP-1 and MASP-1 or -3. We investigated this in NHS, citrated plasma, heparinized plasma, and EDTA plasma by applying the sandwich ELISA assay also used for detection of heterocomplexes. All sample types gave a dose-dependent dilution pattern (Fig. 8), showing that MAP-1 and MASP-1 and/or -3 are found circulating together in some form of complexes. However, with this assay, it is impossible to predict what kind of complexes we detect, as this setup will detect both free circulating heterocomplexes and complexes of different compositions bound to PRMs. Apparently EDTA plasma still contains MASP complexes, which is in agreement with previous observations stating that a high salt concentration is needed in addition to EDTA to dissociate PRM/MASP complexes (25). In vivo MAP-1 heterocomplexes in depleted ficolin-3–deficient human serum

To expose whether MAP-1 and MASPs are directly bound, we 2 2 2 2 FIGURE 9. Depletion of FCN3 / serum. FCN3 / serum was de- focused on the detection of potential free circulating heterocom- pleted (Depl.) for ficolin-1, ficolin-2, MBL, and CL-11. Ficolin-1 (A), Downloaded from plexes to avoid intermediary linkages. This required a depletion of ficolin-2 (B), ficolin-3 (C), and MBL (D) levels were measured in sandwich the MASPs associated with PRMs from serum. However, the ELISA assays. CL-11 depletion was shown in Western blotting (E). possibility of codepleting almost all MASPs would be high in NHS. Thus, we used serum from an FCN32/2 patient to prevent the otherwise large bulk of ficolin-3 from pulling out bound MASPs. have been reported so far (12, 14–16). Yet, it has been shown Fig. 9 shows the successful depletion of ficolin-1, -2, MBL, and that the homologs classical pathway serine proteases C1r and C1s

CL-11, and the complex measurements from nondepleted and form heteromers and that C1s is able to engage in both hetero- http://www.jimmunol.org/ depleted ficolin-3–deficient serum are illustrated in Fig. 10 using tetrameric units with C1r as well as in homodimers at physio- the same sandwich ELISA assay as before. Fig. 10A shows the logical calcium conditions (28). This suggests a more flexible nondepleted FCN32/2 serum, which resembles previous results situation in terms of hetero-oligomerization, and the composition from NHS and plasma. The result after depletion is shown in complexity for the lectin pathway–associated serine proteases and Fig. 10B, and despite the depletion of ficolins, MBL, and CLs, we nonenzymatic proteins might be much greater than previously could still detect complexes between MAP-1 and MASP-1 or -3, perceived. Therefore, we decided to investigate whether the three implying that free circulating heterocomplexes were present in the different variants derived from the MASP1 gene could engage in serum. We do, however, see a significant decline in the signal from hetero-oligomerization, suggesting both a novel mechanism for the depleted serum, suggesting that the major part of the MASP MAP-1 inhibition of complement activation and also a putative by guest on October 1, 2021 fraction is associated in larger PRM complexes, which has also different mode of cross-activation between the MASPs. previously been shown (13). Finally, we were able to eliminate the Our data indicate that rMAP-1 is able to assemble with rMASP-1 residual complex signal using an MAP-1/MASP-1/-3 depletion Ab and rMASP-3 in heterocomplexes under physiological calcium (Fig. 10B). A schematic model of the PRM-bound MASP com- conditions. Also MASP-1 and MASP-3 form heterocomplexes in plexes is shown in Fig. 11. a calcium-dependent manner. The heterocomplex measurements in ELISA did not distinguish between heterodimers, -tetramers, or Discussion even higher oligomers. However, SEC showed that the MAP-1 It is uncertain whether MASPs, MAP-1, and small MBL-associated heterocomplex solutions generated a new peak in between the protein only form homodimers or whether they can interact MASP-1 or -3 and MAP-1 dimer peaks, corresponding to a het- crosswise the various molecules. MASP-1 and the nonproteolytic erodimer, and the depicted Western blots showed that calcium- MASP2 variant small MBL-associated protein (also known as dependent heterocomplexes were indeed formed. The earliest Map19) have been mentioned as possible binding partners (26, visible peaks in the SEC spectrograms were impurities and not 27), but besides that, only homodimers of MASPs and MAP-1 higher oligomers or unspecific MASP-derived aggregates, which supports the proposed formation of a specific calcium-dependent heterocomplex. It moreover supports previous characterizations of MAP-1 or the MASPs that do not report on oligomeric forms higher than dimers (14, 15, 29). The Western blots of the MAP-1–MASP-1 SEC fractions revealed an interesting difference in the band pattern, as activated MASP-1 was dominant in the EDTA-treated samples as opposed to the non–EDTA treated. This difference could be due to inhibition properties of MAP-1 on MASP-1 activation when the molecules are assembled in the calcium buffer. It can be argued that the recombinant expressed MASP-1 has been without MAP-1 prior to the experiment and could have undergone autoactivation FIGURE 8. Complexes comprising MAP-1 and MASP-1 and/or -3 hu- at that stage, thus excluding MAP-1 as the decisive factor. How- man serum and plasma. Serial dilutions of NHS, citrated plasma, heparinized plasma, and EDTA plasma were analyzed regarding the content of ever, it is possible that MAP-1 preferably binds to the nonactivated complexes containing MAP-1 and MASP-1/-3. Results are presented as MASP-1 form and that dissociation by EDTA instantly causes means of triple determinations 6 SD. MASP-1 to autoactivate, suggesting that MAP-1 has an inhibitory 4358 HETEROCOMPLEXES IN THE LECTIN COMPLEMENT PATHWAY

cellular setting encountering both secreted rMAP-1 and rMASP-3. This resulted in the formation of heterocomplexes, and the relative levels of heterocomplexes and MAP-1 homodimers in the culture supernatant appeared to be approximately equivalent. There were no PRMs present in the culturing media; thus, MAP-1 and MASP- 3 bound directly to each other, and the proteins appear to have an appreciable tendency to gather in heterocomplexes after being secreted into the extracellular environment. SEC fractions of the coculture supernatant furthermore showed that MAP-1 and MASP-3 most likely associate in dimers. We elucidated the functional aspect of heterodimers, showing that they were capable of binding to ficolin-3 in solution. This means that ficolin-3 pos- sibly has the ability to carry heterodimers as well as homodimers in the blood circulation. Our measurements in NHS and various plasma types showed that complexes in general containing both MAP-1 and MASPs are found in the circulation. It did, however, not lift the veil on how these complexes actually assemble. It is difficult to examine whether MASPs are directly bound in heterocomplexes, because the PRMs mask the composition of the associated MASPs. Con- Downloaded from sequently, we depleted ficolin-3–deficient serum for the four FIGURE 10. Free heterocomplexes in depleted FCN32/2 serum. remaining PRMs ficolin-1, -2, MBL, and CL-11, and by use of this FCN32/2 serum both nondepleted (A) and depleted (B) was applied in the method, we observed that free circulating heterocomplexes of ELISA setup, showing that nondepleted serum contained complexes MAP-1 and MASP-1 or -3 could in fact be measured in serum. It comprising MAP-1 and MASP-1/-3. The depleted serum revealed that it is, moreover, possible that the actual heterocomplex level is higher was possible to detect free circulating heterocomplexes, namely MAP-1 than we detected, because this assay excluded the majority of http://www.jimmunol.org/ and MASP-1 or -3, bound directly to each other. Results are presented as MASPs through codepletion with the PRMs. means of triple determinations 6 SD. Affinity purification of MAP-1 from human plasma and sub- sequent mass spectrometric peptide mapping and sequencing role in these heterocomplexes. MASP-3 do not autoactivate (30), analysis showed that MBL, ficolin-2, ficolin-3, CL-11, as well as which explains the absent MASP-3 a-chain bands. MASP-3 has MASP-1, MASP-2, and MASP-3 were copurified with MAP-1 also been mentioned as a possible inhibitor of the LCP (19), and in (A. Rosbjerg, L. Munthe-Fog, P. Garred, and M.-O. Skjoedt, the MASP-1–MASP-3 in vitro heterocomplex, we see that MASP- unpublished observations). We are aware that the novel CL CL- 1 is present in the full-length and not the activated form (Fig. 2). 10 (also known as CL-L1) could be a possible candidate in the by guest on October 1, 2021 Based on these findings, we proceeded to a more physiologically family comprising circulating MASP binding partners (31); relevant setting by investigating heterocomplex formation in cell however, we assume that depletion of CL-11 also will include culture, thus avoiding the EDTA-mediated dissociation step and depletion of CL-10 because both CLs reportedly bind to mannose more closely resembling an in vivo situation in which Ca2+ is (32). Moreover, it has recently been shown that CL-10 and CL-11 always present in the extracellular space. We explored two pos- peptide chains presumably form complexes, thus depletion of one sibilities of complex formation: intracellular and extracellular of the molecules will lead to depletion of the other (33). It should association of MAP-1 and MASP-3. Cotransfection of CHO cells be mentioned that depletion was verified in ELISA assays using resembled the intracellular situation and resulted in hetero- lower serum concentrations than the concentration applied for complex formation between the coexpressed proteins MAP-1 and complex measurements because higher serum concentrations MASP-3. This suggests that heterocomplexes may form in the would exceed the limit for these assays. secretory pathway prior to protein secretion. The extracellular Combined, these data suggest that the MASPs have the ability complex formation was assessed by cocultivating the two different to assemble in heterocomplexes and in that way contribute to a populations of transfected CHO cells, thereby creating an extra- cross activation, as suggested by several recent papers (34–37). In

FIGURE 11. Schematic model of hypothesized complexes in the LCP. MBL/ficolin/CL-11 and MASPs/MAP-1 can potentially associate in various ways in which one does not automatically exclude the other. In this figure, we show a schematic overview of different combinations including both MASP homodimers (red/red or green/green protease domains in one dimer) and heterodimers (red/green). The proteolytic coactivation between MASPs can, according to these models, happen through either cis-ortrans-activation. We have defined cis-activation as a mechanism happening between the protease domains within the same dimer and trans-activation as a mechanism happening across two dimers. The Journal of Immunology 4359 contrast, heterocomplexes containing MAP-1 might regulate the at least two different possible ways by which MAP-1 can inhibit protease activity not only by molecular displacement but also by the LCP, through direct competition with the MASPs in PRM steric hindrance on the same protein complex. binding and by engaging in heterocomplexes with the MASPs. We have not quantified the level of heterocomplexes in serum Note added in proof. During the processing of this work, Pare´j compared with homodimers, because the PRM linkage makes this et al. (39) published observations that are consistent with the pres- extremely challenging. It is therefore not possible to conclude on ent results. the true significance of these complexes in the blood. Yet, serum levels are perhaps not so relevant considering the tissue distri- Disclosures bution of MAP-1. MAP-1 is likely to function as a local regulator, The authors have no financial conflicts of interest. with high expression levels in heart and skeletal muscle (10). Heterocomplex formation could thus have an impact on LCP activation at these sites and local inflammatory conditions [e.g., References changing pH could perhaps affect the constellation of MASP 1. Garred, P., C. Honore´, Y. J. Ma, L. Munthe-Fog, and T. Hummelshøj. 2009. complexes, considering recent findings that pH influences the MBL2, FCN1, FCN2 and FCN3-The genes behind the initiation of the lectin pathway of complement. Mol. Immunol. 46: 2737–2744. binding between the two PRMs ficolin-1 and PTX3 (38)]. 2. Ma, Y. J., M.-O. Skjoedt, and P. Garred. 2013. Collectin-11/MASP complex Even though dissociation of the generated in vitro homodimers formation triggers activation of the lectin complement pathway—the fifth lectin was inflicted artificially in this study, it may resemble a physio- pathway initiation complex. J. Innate Immun. 5: 242–250. 3. Collard, C. D., A. Va¨keva¨, M. A. Morrissey, A. Agah, S. A. Rollins, logical process in which MASPs are secreted as monomers and then W. R. Reenstra, J. A. Buras, S. Meri, and G. L. Stahl. 2000. Complement ac- face a many-fold increased calcium level outside the cell, forcing tivation after oxidative stress: role of the lectin complement pathway. Am. J. a dimerization. MASPs could also be secreted as dimers for which Pathol. 156: 1549–1556. Downloaded from 4. Walsh, M. C., T. Bourcier, K. Takahashi, L. Shi, M. N. Busche, R. P. Rother, coexpression of different MASPs in the same cell is a prerequisite S. D. Solomon, R. A. Ezekowitz, and G. L. Stahl. 2005. Mannose-binding lectin for heterodimer formation. is a regulator of inflammation that accompanies myocardial ischemia and reperfusion injury. J. Immunol. 175: 541–546. We mimicked both scenarios in cell-based experiments. Re- 5. Hart, M. L., K. A. Ceonzo, L. A. Shaffer, K. Takahashi, R. P. Rother, sembling the first situation, we observed that coculturing of MAP-1 W. R. Reenstra, J. A. Buras, and G. L. Stahl. 2005. Gastrointestinal ischemia- and MASP-3 CHO cells mediated a profound heterocomplex reperfusion injury is lectin complement pathway dependent without involving J. Immunol. C1q. 174: 6373–6380. http://www.jimmunol.org/ formation, meaning that monomers could associate after secretion 6. de Vries, B., S. J. Walter, C. J. Peutz-Kootstra, T. G. Wolfs, L. W. van Heurn, and into the extracellular space. Resembling the second situation, we W. A. Buurman. 2004. The mannose-binding lectin-pathway is involved in showed that coexpression also conducted heterocomplex forma- complement activation in the course of renal ischemia-reperfusion injury. Am. J. Pathol. 165: 1677–1688. tion, which is in line with a previous study by Degn and colleagues 7. Møller-Kristensen, M., W. Wang, M. Ruseva, S. Thiel, S. Nielsen, K. Takahashi, (18). However, based on our coculture experiments, it is possible L. Shi, A. Ezekowitz, J. C. Jensenius, and M. Gadjeva. 2005. Mannan-binding lectin recognizes structures on ischaemic reperfused mouse kidneys and is im- that what was initially perceived as intracellular complex forma- plicated in tissue injury. Scand. J. Immunol. 61: 426–434. tion may also be a postsecretory phenomenon, because our co- 8. Morrison, H., J. Frye, G. Davis-Gorman, J. Funk, P. McDonagh, G. Stahl, and culture experiments indeed indicate that postsecretory assembly of L. Ritter. 2011. The contribution of mannose binding lectin to reperfusion injury after ischemic stroke. Curr. Neurovasc. Res. 8: 52–63. both homodimers and heterodimers takes place.

9. Schwaeble, W. J., N. J. Lynch, J. E. Clark, M. Marber, N. J. Samani, Y. M. Ali, by guest on October 1, 2021 Reflecting on the in vivo processes, we have previously seen in T. Dudler, B. Parent, K. Lhotta, R. Wallis, et al. 2011. Targeting of mannan- immunohistochemical staining that MAP-1 is synthesized uni- binding lectin-associated serine protease-2 confers protection from myocardial and gastrointestinal ischemia/reperfusion injury. Proc. Natl. Acad. Sci. USA 108: formly in liver hepatocytes (10), and because MASP transcription 7523–7528. is located in the liver as well (10, 26), it is plausible to envisage 10. Skjoedt, M.-O., T. Hummelshoj, Y. Palarasah, C. Honore, C. Koch, K. Skjodt, and P. Garred. 2010. A novel mannose-binding lectin/ficolin-associated protein that MAP-1 and the MASPs are synthesized simultaneously in the is highly expressed in heart and skeletal muscle tissues and inhibits complement same cells through alternative splicing. Future studies are needed activation. J. Biol. Chem. 285: 8234–8243. to clarify this field. 11.Pavlov,V.I.,M.-O.Skjoedt,Y.SiowTan,A.Rosbjerg,P.Garred,and G. L. Stahl. 2012. Endogenous and natural complement inhibitor attenuates Nevertheless, we can imagine a physiological setting in which myocardial injury and arterial thrombogenesis. Circulation 126: 2227– different MASPs gather in heterocomplexes and convey either 2235. complement activation or inhibition. It has recently been suggested 12. Teillet, F., C. Gaboriaud, M. Lacroix, L. Martin, G. J. Arlaud, and N. M. Thielens. 2008. Crystal structure of the CUB1-EGF-CUB2 domain of that two different MASP homodimers can engage in a common human MASP-1/3 and identification of its interaction sites with mannan-binding complex with one PRM (18). This is not necessarily in opposition lectin and ficolins. J. Biol. Chem. 283: 25715–25724. 13. Skjoedt, M.-O., T. Hummelshoj, Y. Palarasah, E. Hein, L. Munthe-Fog, C. Koch, to our observations, as the two scenarios are not mutually exclu- K. Skjodt, and P. Garred. 2011. Serum concentration and interaction properties sive. It is not unlikely that varieties of complexes occur. Lower of MBL/ficolin associated protein-1. Immunobiology 216: 625–632. oligomers of PRMs might not be able to hold two homodimers, 14. Skjoedt, M.-O., P. Roversi, T. Hummelshøj, Y. Palarasah, A. Rosbjerg, S. Johnson, S. M. Lea, and P. Garred. 2012. Crystal structure and func- but instead favor binding of heterodimers. tional characterization of the complement regulator mannose-binding lectin Taken together, these findings suggest a high degree of com- (MBL)/ficolin-associated protein-1 (MAP-1). J. Biol. Chem. 287: 32913– plexity of the activation and regulation of the LCP with multiple 32921. 15. Chen, C. B., and R. Wallis. 2001. Stoichiometry of complexes between PRMs and a range of possible combinations of associated serine mannose-binding protein and its associated serine proteases. Defining functional proteases and nonenzymatic proteins. At this point, we can set up a units for complement activation. J. Biol. Chem. 276: 25894–25902. 16. Thielens, N. M., S. Cseh, S. Thiel, T. Vorup-Jensen, V. Rossi, J. C. Jensenius, and range of possible LCP complex models, as illustrated in Fig. 11; one G. J. Arlaud. 2001. Interaction properties of human mannan-binding lectin can envision that an exchange of an MASP homodimer into a het- (MBL)-associated serine proteases-1 and -2, MBL-associated protein 19, and erodimer containing MAP-1 (not shown) would inhibit activation of MBL. J. Immunol. 166: 5068–5077. 17. Gingras, A. R., U. V. Girija, A. H. Keeble, R. Panchal, D. A. Mitchell, the LCP because MAP-1 lacks the serine protease domain, thereby P. C. Moody, and R. Wallis. 2011. Structural basis of mannan-binding lectin causing a disruption of the mutual coactivation of MASPs. recognition by its associated serine protease MASP-1: implications for com- Additional studies on the stoichiometry and distribution of the plement activation. Structure 19: 1635–1643. 18. Degn, S. E., L. Jensen, T. Olszowski, J. C. Jensenius, and S. Thiel. 2013. Co- different MASP complexes might clarify these aspects further, but complexes of MASP-1 and MASP-2 associated with the soluble pattern- heterocomplexes might be a part of the explanation as to how recognition molecules drive lectin pathway activation in a manner inhibitable by MAp44. J. Immunol. 191: 1334–1345. cross-activation occurs between the different MASPs in cis-or 19. Skjoedt, M.-O., Y. Palarasah, L. Munthe-Fog, Y. Jie Ma, G. Weiss, K. Skjodt, trans-activation modes. As a consequence, this also gives rise to C. Koch, and P. Garred. 2010. MBL-associated serine protease-3 circulates in 4360 HETEROCOMPLEXES IN THE LECTIN COMPLEMENT PATHWAY

high serum concentrations predominantly in complex with Ficolin-3 and reg- protease (MASP)-3 suggests an activation mechanism different from that of ulates Ficolin-3 mediated complement activation. Immunobiology 215: 921–931. MASP-1 and MASP-2. J. Immunol. 172: 4342–4350. 20. Munthe-Fog, L., T. Hummelshoj, C. Honore´, M. E. Moller, M. O. Skjoedt, 31. Axelgaard, E., L. Jensen, T. F. Dyrlund, H. J. Nielsen, J. J. Enghild, S. Thiel, and I. Palsgaard, N. Borregaard, H. O. Madsen, and P. Garred. 2012. Variation in J. C. Jensenius. 2013. Investigations on collectin liver 1. J. Biol. Chem. 288: FCN1 affects biosynthesis of ficolin-1 and is associated with outcome of sys- 23407–23420. temic inflammation. Genes Immun. 13: 515–522. 32. Selman, L., and S. Hansen. 2012. Structure and function of collectin liver 1 21. Munthe-Fog, L., T. Hummelshøj, B. E. Hansen, C. Koch, H. O. Madsen, (CL-L1) and collectin 11 (CL-11, CL-K1). Immunobiology 217: 851–863. K. Skjødt, and P. Garred. 2007. The impact of FCN2 polymorphisms and hap- 33. Henriksen, M. L., J. Brandt, J.-P. Andrieu, C. Nielsen, P. H. Jensen, lotypes on the Ficolin-2 serum levels. Scand. J. Immunol. 65: 383–392. U. Holmskov, T. J. D. Jorgensen, Y. Palarasah, N. M. Thielens, and S. Hansen. 22. Munthe-Fog, L., T. Hummelshøj, Y. J. Ma, B. E. Hansen, C. Koch, 2013. Heteromeric complexes of native collectin kidney 1 and collectin liver 1 H. O. Madsen, K. Skjødt, and P. Garred. 2008. Characterization of a polymor- are found in the circulation with MASPs and activate the complement system. J. phism in the coding sequence of FCN3 resulting in a Ficolin-3 (Hakata antigen) Immunol. 191: 6117–6127. deficiency state. Mol. Immunol. 45: 2660–2666. 34. Kocsis, A., K. A. Ke´kesi, R. Sza´sz, B. M. Ve´gh, J. Balczer, J. Dobo´, 23. Larsen, F., H. O. Madsen, R. B. Sim, C. Koch, and P. Garred. 2004. Disease- P. Za´vodszky, P. Ga´l, and G. Pa´l. 2010. Selective inhibition of the lectin pathway associated mutations in human mannose-binding lectin compromise oligomeri- of complement with phage display selected peptides against mannose-binding zation and activity of the final protein. J. Biol. Chem. 279: 21302–21311. lectin-associated serine protease (MASP)-1 and -2: significant contribution of 24. Munthe-Fog, L., T. Hummelshøj, C. Honore´, H. O. Madsen, H. Permin, and MASP-1 to lectin pathway activation. J. Immunol. 185: 4169–4178. P. Garred. 2009. Immunodeficiency associated with FCN3 mutation and ficolin-3 35. He´ja, D., A. Kocsis, J. Dobo´, K. Szila´gyi, R. Sza´sz, P. Za´vodszky, G. Pa´l, and deficiency. N. Engl. J. Med. 360: 2637–2644. P. Ga´l. 2012. Revised mechanism of complement lectin-pathway activation re- 25. Thiel, S., S. V. Petersen, T. Vorup-Jensen, M. Matsushita, T. Fujita, C. M. Stover, vealing the role of serine protease MASP-1 as the exclusive activator of MASP- W. J. Schwaeble, and J. C. Jensenius. 2000. Interaction of C1q and mannan- 2. Proc. Natl. Acad. Sci. USA 109: 10498–10503. binding lectin (MBL) with C1r, C1s, MBL-associated serine proteases 1 and 2, 36. Megyeri, M., V. Harmat, B. Major, A´ .Ve´gh, J. Balczer, D. He´ja, K. Szila´gyi, and the MBL-associated protein MAp19. J. Immunol. 165: 878–887. D. Datz, G. Pa´l, P. Za´vodszky, et al. 2013. Quantitative characterization of the 26. Takahashi, M., Y. Endo, T. Fujita, and M. Matsushita. 1999. A truncated form of activation steps of mannan-binding lectin (MBL)-associated serine proteases mannose-binding lectin-associated serine protease (MASP)-2 expressed by al- (MASPs) points to the central role of MASP-1 in the initiation of the comple- ternative polyadenylation is a component of the lectin complement pathway. Int. ment lectin pathway. J. Biol. Chem. 288: 8922–8934. Immunol. 11: 859–863. 37. Degn, S. E., L. Jensen, A. G. Hansen, D. Duman, M. Tekin, J. C. Jensenius, and Downloaded from 27. Matsushita, M., S. Thiel, J. C. Jensenius, I. Terai, and T. Fujita. 2000. Proteolytic S. Thiel. 2012. Mannan-binding lectin-associated serine protease (MASP)-1 is activities of two types of mannose-binding lectin-associated serine protease. J. crucial for lectin pathway activation in human serum, whereas neither MASP-1 Immunol. 165: 2637–2642. nor MASP-3 is required for alternative pathway function. J. Immunol. 189: 28. Weiss, V., C. Fauser, and J. Engel. 1986. Functional model of subcomponent C1 3957–3969. of human complement. J. Mol. Biol. 189: 573–581. 38.Ma,Y.J.,A.Doni,L.Romani,H.J.Jurgensen,€ N. Behrendt, A. Mantovani, 29. Teillet, F., B. Dublet, J. P. Andrieu, C. Gaboriaud, G. J. Arlaud, and and P. Garred. 2013. Ficolin-1-PTX3 complex formation promotes clearance N. M. Thielens. 2005. The two major oligomeric forms of human mannan- of altered self-cells and modulates IL-8 production. J. Immunol. 191: 1324–

binding lectin: chemical characterization, carbohydrate-binding properties, and 1333. http://www.jimmunol.org/ interaction with MBL-associated serine proteases. J. Immunol. 174: 2870–2877. 39. Pare´j, K., A´ . Hermann, N. Dona´th, P. Za´vodsky, P. Ga´l, and J. Dobo´. 2014. 30. Zundel, S., S. Cseh, M. Lacroix, M. R. Dahl, M. Matsushita, J. P. Andrieu, Dissociation and re-association studies on the interaction domains of mannan- W. J. Schwaeble, J. C. Jensenius, T. Fujita, G. J. Arlaud, and N. M. Thielens. binding lectin (MBL)-associated serine proteases, MASP-1 and MASP-2, pro- 2004. Characterization of recombinant mannan-binding lectin-associated serine vide evidence for heterodimer formation. Mol. Immunol. 59: 1–9. by guest on October 1, 2021