- Email: [email protected]
Adiponectin binds C1q and activates the classical pathway of complement
Available online at www.sciencedirect.com
Biochemical and Biophysical Research Communications 367 (2008) 560–565 www.elsevier.com/locate/ybbrc
Adiponectin binds C1q and activates the classical pathway of complement Philip W. Peake *, Yvonne Shen, Alexandra Walther, John A. Charlesworth Division of Medicine, Prince of Wales Hospital, High Street, Randwick, Sydney, NSW 2031, Australia Received 20 December 2007 Available online 7 January 2008
Abstract The adipose-specific protein adiponectin binds to a number of target molecules, including damaged endothelium and the surface of apoptotic cells. However, the significance of this binding remains unclear. This study demonstrates the binding of purified C1q to recombinant adiponectin under physiological conditions, and the dependence of this upon Ca++ and Mg++. Binding was enhanced by metaperiodate-mediated destruction of glucosylgalactosyl sugars on adiponectin. Adiponectin was bound by the globular domain of the A chain of collagenase-digested C1q, and C1q binding induced deposition of C4 and C3 through activation of the classical complement pathway. After Western blotting, affinity-purified adiponectin from human serum bound C1q, whereas adiponectin in whole serum did not, unless pre-treated with metaperiodate. These results suggest adiponectin is member of the pattern-recognition family of defence collagens, able to bind target molecules and activate complement. It may therefore play an important role in innate immunity and autoimmune phenomena. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Adiponectin; C3; C4; C1q; Complement; Glucosylgalactosyl
The adipose-specific glycoprotein adiponectin circulates in high concentration, and has an important role in insulin sensitivity and energy homeostasis [1,2]. It also displays anti-inflammatory and anti-atherogenic properties. Supplementation in vivo ameliorates atherosclerosis in apo-E deficient mice [3], and protects the heart from ischemia– reperfusion injury [4]. Adiponectin-deficient mice show increased neointimal thickening and proliferation of vascular smooth muscle cells in mechanically injured arteries [5], and adiponectin infiltrates damaged endothelium, binding to vascular matrix proteins such as collagens I, III, and IV [6]. Lipopolysaccharide and a number of growth factors are bound by adiponectin, thus controlling their bioavailability and inflammatory potential at a pre-receptor level [7–9]. Adiponectin also binds to early apoptotic cells and is important in their clearance by macrophages [10]. This study demonstrates the binding of C1q by both recombinant and native human adiponectin. Divalent cat-
*
Corresponding author. Fax: +61 02 9382 4409. E-mail address: [email protected] (P.W. Peake).
0006-291X/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2007.12.161
ions and the presence of intact glucosylgalactosyl sugars were important in controlling this interaction, which led to the activation of the classical pathway of complement. Both the adiponectin-mediated binding of C1q and the deposition of complement fragments on target molecules could lead to the uptake of these molecules by specific receptors. Methods Adiponectin Carrier-free recombinant human adiponectin was purchased from R&D [Minneapolis, MN]. Elisa assays The interaction of adiponectin with bound C1q. C1q at 10 lg/ml in phosphate buffered saline [PBS] was coated on Nunc Maxisorp wells and non-specific binding sites quenched by incubation with 3% bovine albumin. Adiponectin was added for 60 min at room temperature and, after washing, bound adiponectin was quantitated by the addition of a biotinylated monoclonal antibody to adiponectin [R&D, Minneapolis, MN,
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
Results Characterization of adiponectin–C1q interaction The binding of C1q to recombinant adiponectin was dose-responsive and saturable, as was the binding of recombinant adiponectin to C1q [Fig. 1]. In both cases there was little binding to wells coated with albumin alone [OD 0.02]. The role of divalent cations and glycosylation The influence of cations and chelating agents in the interaction between adiponectin and C1q was determined by their effect on the binding of C1q to recombinant adiponectin coated to Elisa wells [Fig. 2A]. When adiponectin was pre-treated with EGTA, specific largely for Ca++, there was little effect, whereas MgEGTA led to a significant decline in the binding of C1q. Mg++ alone had a similar, although lesser effect, while Ca++ had no significant effect on binding. As adiponectin is a glycoprotein, we also investigated the role of sugar residues on its interaction with C1q. Pre-treatment of adiponectin with 10 mM metaperiodate
C1q binding [OD450]
1.5
1.0
0.5
0.0 0
2
4
6
8
10
C1q [ug/ml]
2.0
Adiponectin binding [OD450]
USA], followed by avidin-peroxidase and tetramethylbenzidine substrate. OD450 readings were taken after the addition of sulphuric acid. The interaction of C1q with bound adiponectin. Adiponectin was coated at 10 lg/ml in PBS. After incubation with 3% bovine albumin, C1q was added for 60 min at room temperature. Bound C1q was quantitated by the addition of a polyclonal antibody specific for C1q [Dako, Glostrup, Denmark], followed by a peroxidase-conjugated monoclonal anti-rabbit IgG [Sigma, Australia]. The effect of various cations and chelators on this interaction was determined by their preincubation for 60 min with bound adiponectin followed by washing and incubation with C1q. In other experiments, the same cations and chelators were preincubated with C1q before addition of the mixture to the Elisa well. The concentration of C1q chosen for evaluating the effect of inhibitors of the interaction of adiponectin and C1q was within the dose-responsive region of the interaction. The effect of the destruction of vicinyl hydroxide-containing sugars was investigated by incubating adiponectin for 30 min in the dark with either 0.1 M acetate buffer, pH 5.5, or with the same buffer containing 1 or 10 mM metaperiodate. Metaperiodate at 1 or 10 mM is specific for sialic acid or all sugars with vicinyl hydroxides, respectively. This treatment did not alter the amount of adiponectin bound to Elisa wells. To investigate the binding of adiponectin from serum to C1q, wells were coated as before with C1q, and non-specific sites blocked with albumin. Serum was added for 60 min, and bound adiponectin detected as described. In some experiments, serum was preincubated with three volumes of 0.1 M acetate buffer, or three volumes of acetate buffer containing 40 mM periodate for 30 min in the dark. Samples were then bufferexchanged via a G25 column [Pharmacia] into PBS before dilution and addition to wells coated with C1q. Complement activation. Non-specific binding sites in wells coated with adiponectin were blocked with human albumin [Sigma]. Fresh serum from a healthy volunteer was diluted in complement fixation diluent [2.8 mM barbituric acid, 145.5 mM NaCl, 0.8 mM MgCl2, 0.3 mM CaCl2, 0.9 mM sodium barbital, pH 7.2, CFD, Oxoid, Australia] and incubated with adiponectin for 30 min at 37 °C. After washing, deposition of C4 was determined by incubation with rabbit anti-C4 [Dako], followed by a peroxidase-conjugated monoclonal antirabbit IgG. Bound C3 was determined by the binding of peroxidaseconjugated anti-C3 [Sigma]. Deposition was compared with that observed in wells coated with albumin alone. To determine the complement pathway[s] active, serum was also pre-incubated with 10 mM ethylenediaminetetraacetic acid [EDTA] or ethyleneglycol-bis-N,N’-tetraacetic acid [EGTA]. Biotinylation. Pure C1q and adiponectin were biotinylated with NHSLC biotin following the manufacturer’s instructions [Pierce, Rockford, IL]. C1q binding by serum adiponectin. Adiponectin was immunopurified from normal sera as described [11], then separated by nonreducing SDS–PAGE. Blots were probed with 2 lg/ml of biotinylated C1q, followed by avidin-peroxidase [Calbiochem, Australia], with detection by chemiluminescence. Other tracks were probed with a monoclonal antibody to adiponectin [BD, Sydney, Australia], followed by peroxidase-conjugated rabbit anti-mouse IgG. Blots were reprobed with a monoclonal antibody to confirm that C1q was bound to adiponectin. Adiponectin binding to collagenase-digested C1q. Purified C1q was incubated overnight at 37 °C with a 1/6 dilution of buffer, or 1.5 mg/ml collagenase [Type VII, purified from Clostridium histolyticum, Sigma] in 0.05 M tricine buffer, pH 7.5, containing 10 mM Ca++ and 400 mM NaCl. The result of incubation was analysed by reducing SDS–PAGE electrophoresis followed by Western blotting. Each track of the gel contained 3.3 lg of native or digested C1q, and the resultant blots were probed with biotinylated adiponectin at 0.7 lg/ml. Other tracks were probed with a polyclonal antibody specific for C1q [Dako], followed by a peroxidase-conjugated monoclonal anti-rabbit IgG [Sigma]. Statistical analysis. Where appropriate, data are shown as means ± SEM, and are representative of at least three independent experiments.
561
1.5 1.0 0.5 0.0 0
1
2 3 4 Adiponectin [ug/ml]
5
Fig. 1. (A) The binding of purified C1q to recombinant human adiponectin coated onto Elisa wells. (B) The binding of recombinant human adiponectin to purified C1q coated onto Elisa wells.
562
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
C1q Binding [Proportion of Control]
A 7 6
a
b
5 4 3 2 1
B
1.0
C1q binding [Proportion]
M gE G
T M Ap g+ re + C pr trea M a+ e tm et ap ED + p trea en re tm t M eri T et od E A -tre e ap a G pr a nt er te TA e-t tm io 10 e r da m pre ea nt te M -t tm r e e 1 m pre atm nt M -tr pr ea ent e- tm tre e at nt m en t M gE G TA M g+ + C a+ ED + T EG A TA
0
0.5
native
0.2
0.4
0.6
0.8
NaCl [M]
Capture of Adiponectin [OD450]
C
0.5
0.4
0.3
acetate 1 mM metaperiodate 10 mM metaperiodate
0.2
0.1
0.0 0.000
The interaction of native adiponectin with C1q To ensure that the interaction between C1q and recombinant adiponectin was representative of that occurring in vivo, experiments with adiponectin in human serum were undertaken. There was little capture of adiponectin from normal serum by immobilised C1q. However, pre-treatment of serum with 10 mM metaperiodate led to a significant increase in the capture of adiponectin [Fig. 2C].
metaperiodate
0.0 0.0
in particular led to enhanced binding of C1q. The antigenicity and quantity of metaperiodate-treated adiponectin bound to the Elisa well remained unaltered [not shown]. In a subsequent series of experiments [Fig. 2A], cations and chelators were included with C1q during incubation with adiponectin. In the presence of EGTA or Mg++ alone there was little effect, whereas incubation with MgEGTA led to a significant increase in the binding of C1q to adiponectin. Ca++ alone inhibited binding. A range of concentrations of NaCl were included in the incubation to investigate the nature of the additional binding sites for C1q created by metaperiodate on adiponectin [Fig. 2B]. Despite the much higher binding of C1q to metaperiodate-treated adiponectin, the degree of inhibition of binding by varying concentrations of NaCl was comparable in both cases.
0.005
0.010
0.015
0.020
Specificity of binding of adiponectin to C1q To identify the adiponectin-binding site on C1q, the latter was digested with collagenase, and the resultant peptide chains separated by reducing SDS–PAGE gels and blotting [Fig. 3]. C1q showed a doublet band consistent with the A and B chains [31 and 30 kDa, respectively], and the C chain [26 kDa] [lane 1]. Partial collagenase digestion yielded three additional moieties, consistent with the globular domains of the A, B, and C chains at 19, 17, and 15 kDa as previously reported [12] [lane 2]. Biotinylated adiponectin bound to the A and C chains of native C1q [lane 3], and to the globular domain of the A chain [lane 4]. We also showed that human adiponectin, affinity-purified from serum, bound biotinylated C1q [Fig. 3B]. Equivalent binding of C1q by affinity-purified adiponectin from both high and low MW fractions of serum [separated by size-sorting chromatography] was also observed [not shown]. Activation of complement by adiponectin
Serum Dilution Fig. 2. (A) The effect on the binding of C1q to solid-phase adiponectin of [a] pre-treatment of adiponectin with cations and/or chelating agents, or metaperiodate, or [b] the co-incubation of cations and/or chelating agents with C1q and solid-phase adiponectin. All cations and chelating agents were at 10 mM. (B) The effect of NaCl concentration on the binding of C1q to acetate buffer-treated, or metaperiodate-treated adiponectin coated onto Elisa wells. (C) The capture of adiponectin from human serum by solid-phase C1q. Serum was pre-treated by incubation with acetate buffer, or acetate buffer containing 1 mM or 10 mM metaperiodate. Binding to an albumin blocked control [OD 0.07] was subtracted from each reading.
C1q binding to adiponectin activated the classical pathway of complement [Fig. 4]. Incubation of serum with adiponectin bound to Elisa wells led to the deposition of C4, whereas the deposition of C4 on wells coated only with albumin was much reduced [Fig. 4A]. The presence of EDTA in sera essentially abrogated the deposition of C4 on adiponectin [not shown]. In addition, C3 from serum was deposited on adiponectin in the presence of CFD, containing Ca++ and Mg++, but not in the presence of
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
563
1.0
. 55 .
1
2
3
4
. 33 . 40
170
. 130
17
11
.
.
0.5
adiponectin albumin
.
100
.
72
.
55
.
.
Fig. 3. (A) A Western blot of a reducing SDS–PAGE analysis of C1q which was subject to collagenase digestion. The blot was probed with antibody to C1q or biotinylated adiponectin. MW markers [kDa] are shown at left. Lanes 1 and 2; C1q detected antigenically following incubation with buffer or collagenase, respectively. Lanes 3 and 4; the binding of biotinylated adiponectin to C1q following incubation with buffer or collagenase, respectively. (B) A Western blot of a non-reducing SDS–PAGE analysis of adiponectin affinity-purified from human serum which was probed with biotinylated human C1q. MW markers [kDa] are shown at left.
MgEGTA or of EDTA [Fig. 4B]. Although metaperiodate treatment of adiponectin led to enhanced binding of C1q [see above], it did not significantly alter the deposition of C3 [not shown]. Discussion Adiponectin has a long half-life in the plasma in comparison with other adipokines [13], and is present in high concentration. It is a member of the defence collagen family, which includes such molecules as Mannan Binding Lectin [14], and our data show adiponectin is also able to activate complement and so play an important role in innate immunity and autoimmune phenomena. This ability most likely depends upon a structural change in the molecule on binding to an appropriate substrate, and offers an explanation as to how the binding of adiponectin to early apoptotic cells [10], or to damaged endothelium vascular matrix proteins such as collagens I, III, and IV [6] has biological sequelae. The binding of adiponectin to surface-bound C1q was saturable, and occurred at physiologically relevant adiponectin concentrations under physiological conditions. This was also true for the interaction of C1q with surface-bound
0.0 0.000
0.005 0.010 0.015 Serum Dilution
0.020
1.5
Deposition of C3 [OD450]
24
Deposition of C4 [OD450]
72
CFD EDTA Mg EGTA
1.0
0.5
0.0 0.000
0.005
0.010
0.015
0.020
Serum Dilution Fig. 4. (A) The deposition of C4 from serum diluted in CFD and incubated on Elisa wells coated with adiponectin or blocked with albumin only. (B) The deposition of C3 from serum diluted in CFD, EDTA or MgEGTA and incubated on Elisa wells coated with adiponectin. Deposition of C3 on albumin coated wells has been subtracted.
adiponectin. There were differential effects on binding depending upon whether the cations or chelators were added to adiponectin alone, then removed, or were added during the course of the interaction with C1q. Thus Mg++ or MgEGTA treatment of adiponectin alone was inhibitory whereas Ca++ and EGTA alone were not, suggesting replacement of Ca++ with Mg++ in adiponectin altered its structure. It is known the tertiary structure of adiponectin differs in its Ca++-bound and Ca++-free forms [15]. Similarly, the globular domain of C1q has an exposed Ca++, and this influences the recognition properties of C1q towards such targets as IgG and C-reactive protein [16]. In this case, the presence of Ca++ during the interaction between adiponectin and C1q inhibited binding; this was not observed with the chelation of Ca++ or the addition of Mg++. However, simultaneous chelation of Ca++ and the addition of Mg++ by MgEGTA led to enhancement of the interaction. This may reflect the insertion of Mg++ into the Ca++-binding site of C1q, with consequent alterations in its binding properties, which depend upon the recognition of patterns of charge in target molecules [17].
564
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
Metaperiodate treatment [10 mM] to destroy vicinyl hydroxides on the sugar residues of recombinant adiponectin increased the binding of C1q. However, no increase was observed for 1 mM metaperiodate, which is specific for sialic acid. Since NaCl-induced disruption of the interactions between C1q and native or periodate-treated adiponectin was identical, it is likely that the additional epitopes uncovered by metaperiodate treatment were similar in binding affinity to those of native protein, and were likely to be in the collagenous tail: the glycosylated hydroxylysines are located in this region of adiponectin. These are responsible for multimer formation [18,19], and their modification induced a conformational change in molecular structure [11]. Human adiponectin derived from whole serum did not bind C1q unless treated with metaperiodate. Similarly, affinity-purified serum adiponectin also bound C1q, but this was subsequent to SDS–PAGE and Western blotting. Such behaviour is appropriate, since otherwise interaction of serum adiponectin with C1q would lead to uncontrolled complement activation. Circulating adiponectin is associated with the complement inhibitor factor H [20], which offers another method of preventing uncontrolled complement activation. Factor H also binds such molecules as fibromodulin, thereby controlling the complement activation that occurs when the globular head of C1q binds fibromodulin. The site on fibromodulin of factor H binding was separate to that for the binding of C1q [21], and in unpublished experiments we found purified factor H did not inhibit the interaction of C1q and adiponectin. The binding of C1q to adiponectin led only to classical pathway activation since the absence of Ca++, or both Ca++ and Mg++ prevented significant deposition of C3. The strong deposition of C4 of the classical pathway confirmed this finding. Although metaperiodate treatment of recombinant adiponectin increased the binding of C1q, it did not increase the amount of C3 deposited, suggesting possible stereochemical limitations upon complement activation. C1q must bridge a gap of limited size between adjacent molecules of IgG, for example, before complement activation can occur. Subsequently, C1q is structurally modified, allowing auto-activation of the C1r and C1s proteases of the C1 complex. The fact that interaction of C1q with adiponectin led to the activation of the classical pathway strongly suggests that the globular head of C1q is involved, and we confirmed this by demonstrating that adiponectin bound to the globular head of collagenase-digested C1q. C1q consists of A, B, and C chains and our data showed binding of adiponectin only to the globular head of the A chain, although it bound to both the A and C chains of the whole molecule. The globular domains of the A, B, and C chains of C1q are thought to be autonomous and to have different binding properties; thus the A chain also bound specifically to apoptotic mononucleocytes [22], while charged arginine and histidine residues in the globular B chain are believed key in the C1q–IgG interaction [23].
Collectins, such as Mannan Binding Lectin, bind selectively to carbohydrate structures on microbial surfaces, and the polysaccharide side chains of lipopolysaccharide were shown to be important in their interaction with adiponectin [8]. C1q binds late apoptotic cells, and the pentose sugars of DNA may be an important ligand for this reaction [24]. Adiponectin binds to and opsonises early apoptotic cells, leading to the uptake of apoptotic cells via macrophages [10]. It may directly interact with calreticulin [10], although this was not fully confirmed subsequently [25]. Our data imply that adiponectin might mediate the removal of apoptotic cells through the interaction of adiponectin-bound C1q with calreticulin and CD91 [26], or via complement activation and opsonisation of bound cells [24]. To summarise, our data show adiponectin may not only act as a scavenging anti-inflammatory agent, but may also induce biological sequelae through complement activation. Serum adiponectin required a conformational change resulting from SDS or metaperiodate treatment before binding C1q. We suggest that on binding to an appropriate surface or ligand, adiponectin changes in conformation, triggering C1q binding, which induces activation of the classical complement pathway. SIGN-R1, a surface-bound C-type lectin, also activates the classical pathway after binding Pneumococcal polysaccharide [27]. References [1] N. Kubota, Y. Terauchi, T. Yamauchi, T. Kubota, M. Moroi, J. Matsui, K. Eto, T. Yamashita, J. Kamon, H. Satoh, W. Yano, P. Froguel, R. Nagai, S. Kimura, T. Kadowaki, T. Noda, Disruption of adiponectin causes insulin resistance and neointimal formation, J. Biol. Chem. 277 (2002) 25863–25866. [2] N. Maeda, I. Shimomura, K. Kishida, H. Nishizawa, M. Matsuda, H. Nagaretani, N. Furuyama, H. Kondo, M. Takahashi, Y. Arita, R. Komuro, N. Ouchi, S. Kihara, Y. Tochino, K. Okutomi, M. Horie, S. Takeda, T. Aoyama, T. Funahashi, Y. Matsuzawa, Diet-induced insulin resistance in mice lacking adiponectin/ACRP30, Nat. Med. 8 (2002) 731–737. [3] Y. Okamoto, S. Kihara, N. Ouchi, M. Nishida, Y. Arita, M. Kumada, K. Ohashi, N. Sakai, I. Shimomura, H. Kobayashi, N. Terasaka, T. Inaba, T. Funahashi, Y. Matsuzawa, Adiponectin reduces atherosclerosis in apolipoprotein E-deficient mice, Circulation 106 (2002) 2767–2770. [4] R. Shibata, K. Sato, D.R. Pimentel, Y. Takemura, S. Kihara, K. Ohashi, T. Funahashi, N. Ouchi, K. Walsh, Adiponectin protects against myocardial ischemia–reperfusion injury through AMPK- and COX-2-dependent mechanisms, Nat. Med. 11 (2005) 1096–1103. [5] M. Matsuda, I. Shimomura, M. Sata, Y. Arita, M. Nishida, N. Maeda, M. Kumada, Y. Okamoto, H. Nagaretani, H. Nishizawa, K. Kishida, R. Komuro, N. Ouchi, S. Kihara, R. Nagai, T. Funahashi, Y. Matsuzawa, Role of adiponectin in preventing vascular stenosis. The missing link of adipo-vascular axis, J. Biol. Chem. 277 (2002) 37487–37491. [6] Y. Okamoto, Y. Arita, M. Nishida, M. Muraguchi, N. Ouchi, M. Takahashi, T. Igura, Y. Inui, S. Kihara, T. Nakamura, S. Yamashita, J. Miyagawa, T. Funahashi, Y. Matsuzawa, An adipocyte-derived plasma protein, adiponectin, adheres to injured vascular walls, Horm. Metab. Res. 32 (2000) 47–50. [7] Y. Arita, S. Kihara, N. Ouchi, K. Maeda, H. Kuriyama, Y. Okamoto, M. Kumada, K. Hotta, M. Nishida, M. Takahashi, T.
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Nakamura, I. Shimomura, M. Muraguchi, Y. Ohmoto, T. Funahashi, Y. Matsuzawa, Adipocyte-derived plasma protein adiponectin acts as a platelet-derived growth factor-BB-binding protein and regulates growth factor-induced common postreceptor signal in vascular smooth muscle cell, Circulation 105 (2002) 2893–2898. P.W. Peake, Y. Shen, L.V. Campbell, J.A. Charlesworth, Human adiponectin binds to bacterial lipopolysaccharide, Biochem. Biophys. Res. Commun. 341 (2006) 108–115. Y. Wang, K.S. Lam, J.Y. Xu, G. Lu, L.Y. Xu, G.J. Cooper, A. Xu, Adiponectin inhibits cell proliferation by interacting with several growth factors in an oligomerization dependent manner, J. Biol. Chem. 280 (2005) 18341–18347. Y. Takemura, N. Ouchi, R. Shibata, T. Aprahamian, M.T. Kirber, R.S. Summer, S. Kihara, K. Walsh, Adiponectin modulates inflammatory reactions via calreticulin receptor-dependent clearance of early apoptotic bodies, J. Clin. Invest. 117 (2007) 375–386. P.W. Peake, J.T. Hughes, Y. Shen, J.A. Charlesworth, Glycosylation of human adiponectin affects its conformation and stability, J. Mol. Endocrinol. 39 (2007) 45–52. F.D. McGrath, M.C. Brouwer, G.J. Arlaud, M.R. Daha, C.E. Hack, A. Roos, Evidence that complement protein C1q interacts with Creactive protein through its globular head region, J. Immunol. 176 (2006) 2950–2957. P.W. Peake, A.D. Kriketos, L.V. Campbell, Y. Shen, J.A. Charlesworth, The metabolism of isoforms of human adiponectin: studies in human subjects and in experimental animals, Eur. J. Endocrinol. 153 (2005) 409–417. S.S. Bohlson, D.A. Fraser, A.J. Tenner, Complement proteins C1q and MBL are pattern recognition molecules that signal immediate and long-term protective immune functions, Mol. Immunol. 44 (2007) 33–43. U. Kishore, C. Gaboriaud, P. Waters, A.K. Shrive, T.J. Greenhough, K.B. Reid, R.B. Sim, G.J. Arlaud, C1q and tumor necrosis factor superfamily: modularity and versatility, Trends Immunol. 25 (2004) 551–561. L.T. Roumenina, A.A. Kantardjiev, B.P. Atanasov, P. Waters, M. Gadjeva, K.B. Reid, A. Mantovani, U. Kishore, M.S. Kojouharova, Role of Ca2+ in the electrostatic stability and the functional activity of the globular domain of human C1q, Biochemistry 44 (2005) 14097– 14109. R. Ghai, P. Waters, L.T. Roumenina, M. Gadjeva, M.S. Kojouharova, K.B. Reid, R.B. Sim, U. Kishore, C1q and its growing family, Immunobiology 212 (2007) 253–266.
565
[18] Y. Wang, K.S. Lam, L. Chan, K.W. Chan, J.B. Lam, M.C. Lam, R.C. Hoo, W.W. Mak, G.J. Cooper, A. Xu, Post-translational modifications of the four conserved lysine residues within the collagenous domain of adiponectin are required for the formation of its high molecular weight oligomeric complex, J. Biol. Chem. 281 (2006) 16391–16400. [19] A.A. Richards, T. Stephens, H.K. Charlton, A. Jones, G.A. Macdonald, J.B. Prins, J.P. Whitehead, Adiponectin multimerization is dependent on conserved lysines in the collagenous domain: evidence for regulation of multimerization by alterations in posttranslational modifications, Mol. Endocrinol. 20 (2006) 1673–1687. [20] Y. Wang, L.Y. Xu, K.S. Lam, G. Lu, G.J. Cooper, A. Xu, Proteomic characterization of human serum proteins associated with the fatderived hormone adiponectin, Proteomics 6 (2006) 3862–3870. [21] A. Sjoberg, P. Onnerfjord, M. Morgelin, D. Heinegard, A.M. Blom, The extracellular matrix and inflammation: fibromodulin activates the classical pathway of complement by directly binding C1q, J. Biol. Chem. 280 (2005) 32301–32308. [22] U. Kishore, S.K. Gupta, M.V. Perdikoulis, M.S. Kojouharova, B.C. Urban, K.B. Reid, Modular organization of the carboxyl-terminal, globular head region of human C1q A, B, and C chains, J. Immunol. 171 (2003) 812–820. [23] M.S. Kojouharova, M.G. Gadjeva, I.G. Tsacheva, A. Zlatarova, L.T. Roumenina, M.I. Tchorbadjieva, B.P. Atanasov, P. Waters, B.C. Urban, R.B. Sim, K.B. Reid, U. Kishore, Mutational analyses of the recombinant globular regions of human C1q A, B, and C chains suggest an essential role for arginine and histidine residues in the C1q–IgG interaction, J. Immunol. 172 (2004) 4351–4358. [24] L.A. Trouw, A.M. Blom, P. Gasque, Role of complement and complement regulators in the removal of apoptotic cells, Mol. Immunol. 45 (2008) 1199–1207. [25] K. Duus, R.T. Pagh, U. Holmskov, P. Hojrup, S. Skov, G. Houen, Interaction of calreticulin with CD40 ligand, TRAIL and Fas ligand, Scand. J. Immunol. 66 (2007) 501–507. [26] R.W. Vandivier, C.A. Ogden, V.A. Fadok, P.R. Hoffmann, K.K. Brown, M. Botto, M.J. Walport, J.H. Fisher, P.M. Henson, K.E. Greene, Role of surfactant proteins A, D, and C1q in the clearance of apoptotic cells in vivo and in vitro: calreticulin and CD91 as a common collectin receptor complex, J. Immunol. 169 (2002) 3978–3986. [27] Y.S. Kang, Y. Do, H.K. Lee, S.H. Park, C. Cheong, R.M. Lynch, J.M. Loeffler, R.M. Steinman, C.G. Park, A dominant complement fixation pathway for pneumococcal polysaccharides initiated by SIGN-R1 interacting with C1q, Cell 125 (2006) 47–58.
Biochemical and Biophysical Research Communications 367 (2008) 560–565 www.elsevier.com/locate/ybbrc
Adiponectin binds C1q and activates the classical pathway of complement Philip W. Peake *, Yvonne Shen, Alexandra Walther, John A. Charlesworth Division of Medicine, Prince of Wales Hospital, High Street, Randwick, Sydney, NSW 2031, Australia Received 20 December 2007 Available online 7 January 2008
Abstract The adipose-specific protein adiponectin binds to a number of target molecules, including damaged endothelium and the surface of apoptotic cells. However, the significance of this binding remains unclear. This study demonstrates the binding of purified C1q to recombinant adiponectin under physiological conditions, and the dependence of this upon Ca++ and Mg++. Binding was enhanced by metaperiodate-mediated destruction of glucosylgalactosyl sugars on adiponectin. Adiponectin was bound by the globular domain of the A chain of collagenase-digested C1q, and C1q binding induced deposition of C4 and C3 through activation of the classical complement pathway. After Western blotting, affinity-purified adiponectin from human serum bound C1q, whereas adiponectin in whole serum did not, unless pre-treated with metaperiodate. These results suggest adiponectin is member of the pattern-recognition family of defence collagens, able to bind target molecules and activate complement. It may therefore play an important role in innate immunity and autoimmune phenomena. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Adiponectin; C3; C4; C1q; Complement; Glucosylgalactosyl
The adipose-specific glycoprotein adiponectin circulates in high concentration, and has an important role in insulin sensitivity and energy homeostasis [1,2]. It also displays anti-inflammatory and anti-atherogenic properties. Supplementation in vivo ameliorates atherosclerosis in apo-E deficient mice [3], and protects the heart from ischemia– reperfusion injury [4]. Adiponectin-deficient mice show increased neointimal thickening and proliferation of vascular smooth muscle cells in mechanically injured arteries [5], and adiponectin infiltrates damaged endothelium, binding to vascular matrix proteins such as collagens I, III, and IV [6]. Lipopolysaccharide and a number of growth factors are bound by adiponectin, thus controlling their bioavailability and inflammatory potential at a pre-receptor level [7–9]. Adiponectin also binds to early apoptotic cells and is important in their clearance by macrophages [10]. This study demonstrates the binding of C1q by both recombinant and native human adiponectin. Divalent cat-
*
Corresponding author. Fax: +61 02 9382 4409. E-mail address: [email protected] (P.W. Peake).
0006-291X/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2007.12.161
ions and the presence of intact glucosylgalactosyl sugars were important in controlling this interaction, which led to the activation of the classical pathway of complement. Both the adiponectin-mediated binding of C1q and the deposition of complement fragments on target molecules could lead to the uptake of these molecules by specific receptors. Methods Adiponectin Carrier-free recombinant human adiponectin was purchased from R&D [Minneapolis, MN]. Elisa assays The interaction of adiponectin with bound C1q. C1q at 10 lg/ml in phosphate buffered saline [PBS] was coated on Nunc Maxisorp wells and non-specific binding sites quenched by incubation with 3% bovine albumin. Adiponectin was added for 60 min at room temperature and, after washing, bound adiponectin was quantitated by the addition of a biotinylated monoclonal antibody to adiponectin [R&D, Minneapolis, MN,
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
Results Characterization of adiponectin–C1q interaction The binding of C1q to recombinant adiponectin was dose-responsive and saturable, as was the binding of recombinant adiponectin to C1q [Fig. 1]. In both cases there was little binding to wells coated with albumin alone [OD 0.02]. The role of divalent cations and glycosylation The influence of cations and chelating agents in the interaction between adiponectin and C1q was determined by their effect on the binding of C1q to recombinant adiponectin coated to Elisa wells [Fig. 2A]. When adiponectin was pre-treated with EGTA, specific largely for Ca++, there was little effect, whereas MgEGTA led to a significant decline in the binding of C1q. Mg++ alone had a similar, although lesser effect, while Ca++ had no significant effect on binding. As adiponectin is a glycoprotein, we also investigated the role of sugar residues on its interaction with C1q. Pre-treatment of adiponectin with 10 mM metaperiodate
C1q binding [OD450]
1.5
1.0
0.5
0.0 0
2
4
6
8
10
C1q [ug/ml]
2.0
Adiponectin binding [OD450]
USA], followed by avidin-peroxidase and tetramethylbenzidine substrate. OD450 readings were taken after the addition of sulphuric acid. The interaction of C1q with bound adiponectin. Adiponectin was coated at 10 lg/ml in PBS. After incubation with 3% bovine albumin, C1q was added for 60 min at room temperature. Bound C1q was quantitated by the addition of a polyclonal antibody specific for C1q [Dako, Glostrup, Denmark], followed by a peroxidase-conjugated monoclonal anti-rabbit IgG [Sigma, Australia]. The effect of various cations and chelators on this interaction was determined by their preincubation for 60 min with bound adiponectin followed by washing and incubation with C1q. In other experiments, the same cations and chelators were preincubated with C1q before addition of the mixture to the Elisa well. The concentration of C1q chosen for evaluating the effect of inhibitors of the interaction of adiponectin and C1q was within the dose-responsive region of the interaction. The effect of the destruction of vicinyl hydroxide-containing sugars was investigated by incubating adiponectin for 30 min in the dark with either 0.1 M acetate buffer, pH 5.5, or with the same buffer containing 1 or 10 mM metaperiodate. Metaperiodate at 1 or 10 mM is specific for sialic acid or all sugars with vicinyl hydroxides, respectively. This treatment did not alter the amount of adiponectin bound to Elisa wells. To investigate the binding of adiponectin from serum to C1q, wells were coated as before with C1q, and non-specific sites blocked with albumin. Serum was added for 60 min, and bound adiponectin detected as described. In some experiments, serum was preincubated with three volumes of 0.1 M acetate buffer, or three volumes of acetate buffer containing 40 mM periodate for 30 min in the dark. Samples were then bufferexchanged via a G25 column [Pharmacia] into PBS before dilution and addition to wells coated with C1q. Complement activation. Non-specific binding sites in wells coated with adiponectin were blocked with human albumin [Sigma]. Fresh serum from a healthy volunteer was diluted in complement fixation diluent [2.8 mM barbituric acid, 145.5 mM NaCl, 0.8 mM MgCl2, 0.3 mM CaCl2, 0.9 mM sodium barbital, pH 7.2, CFD, Oxoid, Australia] and incubated with adiponectin for 30 min at 37 °C. After washing, deposition of C4 was determined by incubation with rabbit anti-C4 [Dako], followed by a peroxidase-conjugated monoclonal antirabbit IgG. Bound C3 was determined by the binding of peroxidaseconjugated anti-C3 [Sigma]. Deposition was compared with that observed in wells coated with albumin alone. To determine the complement pathway[s] active, serum was also pre-incubated with 10 mM ethylenediaminetetraacetic acid [EDTA] or ethyleneglycol-bis-N,N’-tetraacetic acid [EGTA]. Biotinylation. Pure C1q and adiponectin were biotinylated with NHSLC biotin following the manufacturer’s instructions [Pierce, Rockford, IL]. C1q binding by serum adiponectin. Adiponectin was immunopurified from normal sera as described [11], then separated by nonreducing SDS–PAGE. Blots were probed with 2 lg/ml of biotinylated C1q, followed by avidin-peroxidase [Calbiochem, Australia], with detection by chemiluminescence. Other tracks were probed with a monoclonal antibody to adiponectin [BD, Sydney, Australia], followed by peroxidase-conjugated rabbit anti-mouse IgG. Blots were reprobed with a monoclonal antibody to confirm that C1q was bound to adiponectin. Adiponectin binding to collagenase-digested C1q. Purified C1q was incubated overnight at 37 °C with a 1/6 dilution of buffer, or 1.5 mg/ml collagenase [Type VII, purified from Clostridium histolyticum, Sigma] in 0.05 M tricine buffer, pH 7.5, containing 10 mM Ca++ and 400 mM NaCl. The result of incubation was analysed by reducing SDS–PAGE electrophoresis followed by Western blotting. Each track of the gel contained 3.3 lg of native or digested C1q, and the resultant blots were probed with biotinylated adiponectin at 0.7 lg/ml. Other tracks were probed with a polyclonal antibody specific for C1q [Dako], followed by a peroxidase-conjugated monoclonal anti-rabbit IgG [Sigma]. Statistical analysis. Where appropriate, data are shown as means ± SEM, and are representative of at least three independent experiments.
561
1.5 1.0 0.5 0.0 0
1
2 3 4 Adiponectin [ug/ml]
5
Fig. 1. (A) The binding of purified C1q to recombinant human adiponectin coated onto Elisa wells. (B) The binding of recombinant human adiponectin to purified C1q coated onto Elisa wells.
562
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
C1q Binding [Proportion of Control]
A 7 6
a
b
5 4 3 2 1
B
1.0
C1q binding [Proportion]
M gE G
T M Ap g+ re + C pr trea M a+ e tm et ap ED + p trea en re tm t M eri T et od E A -tre e ap a G pr a nt er te TA e-t tm io 10 e r da m pre ea nt te M -t tm r e e 1 m pre atm nt M -tr pr ea ent e- tm tre e at nt m en t M gE G TA M g+ + C a+ ED + T EG A TA
0
0.5
native
0.2
0.4
0.6
0.8
NaCl [M]
Capture of Adiponectin [OD450]
C
0.5
0.4
0.3
acetate 1 mM metaperiodate 10 mM metaperiodate
0.2
0.1
0.0 0.000
The interaction of native adiponectin with C1q To ensure that the interaction between C1q and recombinant adiponectin was representative of that occurring in vivo, experiments with adiponectin in human serum were undertaken. There was little capture of adiponectin from normal serum by immobilised C1q. However, pre-treatment of serum with 10 mM metaperiodate led to a significant increase in the capture of adiponectin [Fig. 2C].
metaperiodate
0.0 0.0
in particular led to enhanced binding of C1q. The antigenicity and quantity of metaperiodate-treated adiponectin bound to the Elisa well remained unaltered [not shown]. In a subsequent series of experiments [Fig. 2A], cations and chelators were included with C1q during incubation with adiponectin. In the presence of EGTA or Mg++ alone there was little effect, whereas incubation with MgEGTA led to a significant increase in the binding of C1q to adiponectin. Ca++ alone inhibited binding. A range of concentrations of NaCl were included in the incubation to investigate the nature of the additional binding sites for C1q created by metaperiodate on adiponectin [Fig. 2B]. Despite the much higher binding of C1q to metaperiodate-treated adiponectin, the degree of inhibition of binding by varying concentrations of NaCl was comparable in both cases.
0.005
0.010
0.015
0.020
Specificity of binding of adiponectin to C1q To identify the adiponectin-binding site on C1q, the latter was digested with collagenase, and the resultant peptide chains separated by reducing SDS–PAGE gels and blotting [Fig. 3]. C1q showed a doublet band consistent with the A and B chains [31 and 30 kDa, respectively], and the C chain [26 kDa] [lane 1]. Partial collagenase digestion yielded three additional moieties, consistent with the globular domains of the A, B, and C chains at 19, 17, and 15 kDa as previously reported [12] [lane 2]. Biotinylated adiponectin bound to the A and C chains of native C1q [lane 3], and to the globular domain of the A chain [lane 4]. We also showed that human adiponectin, affinity-purified from serum, bound biotinylated C1q [Fig. 3B]. Equivalent binding of C1q by affinity-purified adiponectin from both high and low MW fractions of serum [separated by size-sorting chromatography] was also observed [not shown]. Activation of complement by adiponectin
Serum Dilution Fig. 2. (A) The effect on the binding of C1q to solid-phase adiponectin of [a] pre-treatment of adiponectin with cations and/or chelating agents, or metaperiodate, or [b] the co-incubation of cations and/or chelating agents with C1q and solid-phase adiponectin. All cations and chelating agents were at 10 mM. (B) The effect of NaCl concentration on the binding of C1q to acetate buffer-treated, or metaperiodate-treated adiponectin coated onto Elisa wells. (C) The capture of adiponectin from human serum by solid-phase C1q. Serum was pre-treated by incubation with acetate buffer, or acetate buffer containing 1 mM or 10 mM metaperiodate. Binding to an albumin blocked control [OD 0.07] was subtracted from each reading.
C1q binding to adiponectin activated the classical pathway of complement [Fig. 4]. Incubation of serum with adiponectin bound to Elisa wells led to the deposition of C4, whereas the deposition of C4 on wells coated only with albumin was much reduced [Fig. 4A]. The presence of EDTA in sera essentially abrogated the deposition of C4 on adiponectin [not shown]. In addition, C3 from serum was deposited on adiponectin in the presence of CFD, containing Ca++ and Mg++, but not in the presence of
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
563
1.0
. 55 .
1
2
3
4
. 33 . 40
170
. 130
17
11
.
.
0.5
adiponectin albumin
.
100
.
72
.
55
.
.
Fig. 3. (A) A Western blot of a reducing SDS–PAGE analysis of C1q which was subject to collagenase digestion. The blot was probed with antibody to C1q or biotinylated adiponectin. MW markers [kDa] are shown at left. Lanes 1 and 2; C1q detected antigenically following incubation with buffer or collagenase, respectively. Lanes 3 and 4; the binding of biotinylated adiponectin to C1q following incubation with buffer or collagenase, respectively. (B) A Western blot of a non-reducing SDS–PAGE analysis of adiponectin affinity-purified from human serum which was probed with biotinylated human C1q. MW markers [kDa] are shown at left.
MgEGTA or of EDTA [Fig. 4B]. Although metaperiodate treatment of adiponectin led to enhanced binding of C1q [see above], it did not significantly alter the deposition of C3 [not shown]. Discussion Adiponectin has a long half-life in the plasma in comparison with other adipokines [13], and is present in high concentration. It is a member of the defence collagen family, which includes such molecules as Mannan Binding Lectin [14], and our data show adiponectin is also able to activate complement and so play an important role in innate immunity and autoimmune phenomena. This ability most likely depends upon a structural change in the molecule on binding to an appropriate substrate, and offers an explanation as to how the binding of adiponectin to early apoptotic cells [10], or to damaged endothelium vascular matrix proteins such as collagens I, III, and IV [6] has biological sequelae. The binding of adiponectin to surface-bound C1q was saturable, and occurred at physiologically relevant adiponectin concentrations under physiological conditions. This was also true for the interaction of C1q with surface-bound
0.0 0.000
0.005 0.010 0.015 Serum Dilution
0.020
1.5
Deposition of C3 [OD450]
24
Deposition of C4 [OD450]
72
CFD EDTA Mg EGTA
1.0
0.5
0.0 0.000
0.005
0.010
0.015
0.020
Serum Dilution Fig. 4. (A) The deposition of C4 from serum diluted in CFD and incubated on Elisa wells coated with adiponectin or blocked with albumin only. (B) The deposition of C3 from serum diluted in CFD, EDTA or MgEGTA and incubated on Elisa wells coated with adiponectin. Deposition of C3 on albumin coated wells has been subtracted.
adiponectin. There were differential effects on binding depending upon whether the cations or chelators were added to adiponectin alone, then removed, or were added during the course of the interaction with C1q. Thus Mg++ or MgEGTA treatment of adiponectin alone was inhibitory whereas Ca++ and EGTA alone were not, suggesting replacement of Ca++ with Mg++ in adiponectin altered its structure. It is known the tertiary structure of adiponectin differs in its Ca++-bound and Ca++-free forms [15]. Similarly, the globular domain of C1q has an exposed Ca++, and this influences the recognition properties of C1q towards such targets as IgG and C-reactive protein [16]. In this case, the presence of Ca++ during the interaction between adiponectin and C1q inhibited binding; this was not observed with the chelation of Ca++ or the addition of Mg++. However, simultaneous chelation of Ca++ and the addition of Mg++ by MgEGTA led to enhancement of the interaction. This may reflect the insertion of Mg++ into the Ca++-binding site of C1q, with consequent alterations in its binding properties, which depend upon the recognition of patterns of charge in target molecules [17].
564
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
Metaperiodate treatment [10 mM] to destroy vicinyl hydroxides on the sugar residues of recombinant adiponectin increased the binding of C1q. However, no increase was observed for 1 mM metaperiodate, which is specific for sialic acid. Since NaCl-induced disruption of the interactions between C1q and native or periodate-treated adiponectin was identical, it is likely that the additional epitopes uncovered by metaperiodate treatment were similar in binding affinity to those of native protein, and were likely to be in the collagenous tail: the glycosylated hydroxylysines are located in this region of adiponectin. These are responsible for multimer formation [18,19], and their modification induced a conformational change in molecular structure [11]. Human adiponectin derived from whole serum did not bind C1q unless treated with metaperiodate. Similarly, affinity-purified serum adiponectin also bound C1q, but this was subsequent to SDS–PAGE and Western blotting. Such behaviour is appropriate, since otherwise interaction of serum adiponectin with C1q would lead to uncontrolled complement activation. Circulating adiponectin is associated with the complement inhibitor factor H [20], which offers another method of preventing uncontrolled complement activation. Factor H also binds such molecules as fibromodulin, thereby controlling the complement activation that occurs when the globular head of C1q binds fibromodulin. The site on fibromodulin of factor H binding was separate to that for the binding of C1q [21], and in unpublished experiments we found purified factor H did not inhibit the interaction of C1q and adiponectin. The binding of C1q to adiponectin led only to classical pathway activation since the absence of Ca++, or both Ca++ and Mg++ prevented significant deposition of C3. The strong deposition of C4 of the classical pathway confirmed this finding. Although metaperiodate treatment of recombinant adiponectin increased the binding of C1q, it did not increase the amount of C3 deposited, suggesting possible stereochemical limitations upon complement activation. C1q must bridge a gap of limited size between adjacent molecules of IgG, for example, before complement activation can occur. Subsequently, C1q is structurally modified, allowing auto-activation of the C1r and C1s proteases of the C1 complex. The fact that interaction of C1q with adiponectin led to the activation of the classical pathway strongly suggests that the globular head of C1q is involved, and we confirmed this by demonstrating that adiponectin bound to the globular head of collagenase-digested C1q. C1q consists of A, B, and C chains and our data showed binding of adiponectin only to the globular head of the A chain, although it bound to both the A and C chains of the whole molecule. The globular domains of the A, B, and C chains of C1q are thought to be autonomous and to have different binding properties; thus the A chain also bound specifically to apoptotic mononucleocytes [22], while charged arginine and histidine residues in the globular B chain are believed key in the C1q–IgG interaction [23].
Collectins, such as Mannan Binding Lectin, bind selectively to carbohydrate structures on microbial surfaces, and the polysaccharide side chains of lipopolysaccharide were shown to be important in their interaction with adiponectin [8]. C1q binds late apoptotic cells, and the pentose sugars of DNA may be an important ligand for this reaction [24]. Adiponectin binds to and opsonises early apoptotic cells, leading to the uptake of apoptotic cells via macrophages [10]. It may directly interact with calreticulin [10], although this was not fully confirmed subsequently [25]. Our data imply that adiponectin might mediate the removal of apoptotic cells through the interaction of adiponectin-bound C1q with calreticulin and CD91 [26], or via complement activation and opsonisation of bound cells [24]. To summarise, our data show adiponectin may not only act as a scavenging anti-inflammatory agent, but may also induce biological sequelae through complement activation. Serum adiponectin required a conformational change resulting from SDS or metaperiodate treatment before binding C1q. We suggest that on binding to an appropriate surface or ligand, adiponectin changes in conformation, triggering C1q binding, which induces activation of the classical complement pathway. SIGN-R1, a surface-bound C-type lectin, also activates the classical pathway after binding Pneumococcal polysaccharide [27]. References [1] N. Kubota, Y. Terauchi, T. Yamauchi, T. Kubota, M. Moroi, J. Matsui, K. Eto, T. Yamashita, J. Kamon, H. Satoh, W. Yano, P. Froguel, R. Nagai, S. Kimura, T. Kadowaki, T. Noda, Disruption of adiponectin causes insulin resistance and neointimal formation, J. Biol. Chem. 277 (2002) 25863–25866. [2] N. Maeda, I. Shimomura, K. Kishida, H. Nishizawa, M. Matsuda, H. Nagaretani, N. Furuyama, H. Kondo, M. Takahashi, Y. Arita, R. Komuro, N. Ouchi, S. Kihara, Y. Tochino, K. Okutomi, M. Horie, S. Takeda, T. Aoyama, T. Funahashi, Y. Matsuzawa, Diet-induced insulin resistance in mice lacking adiponectin/ACRP30, Nat. Med. 8 (2002) 731–737. [3] Y. Okamoto, S. Kihara, N. Ouchi, M. Nishida, Y. Arita, M. Kumada, K. Ohashi, N. Sakai, I. Shimomura, H. Kobayashi, N. Terasaka, T. Inaba, T. Funahashi, Y. Matsuzawa, Adiponectin reduces atherosclerosis in apolipoprotein E-deficient mice, Circulation 106 (2002) 2767–2770. [4] R. Shibata, K. Sato, D.R. Pimentel, Y. Takemura, S. Kihara, K. Ohashi, T. Funahashi, N. Ouchi, K. Walsh, Adiponectin protects against myocardial ischemia–reperfusion injury through AMPK- and COX-2-dependent mechanisms, Nat. Med. 11 (2005) 1096–1103. [5] M. Matsuda, I. Shimomura, M. Sata, Y. Arita, M. Nishida, N. Maeda, M. Kumada, Y. Okamoto, H. Nagaretani, H. Nishizawa, K. Kishida, R. Komuro, N. Ouchi, S. Kihara, R. Nagai, T. Funahashi, Y. Matsuzawa, Role of adiponectin in preventing vascular stenosis. The missing link of adipo-vascular axis, J. Biol. Chem. 277 (2002) 37487–37491. [6] Y. Okamoto, Y. Arita, M. Nishida, M. Muraguchi, N. Ouchi, M. Takahashi, T. Igura, Y. Inui, S. Kihara, T. Nakamura, S. Yamashita, J. Miyagawa, T. Funahashi, Y. Matsuzawa, An adipocyte-derived plasma protein, adiponectin, adheres to injured vascular walls, Horm. Metab. Res. 32 (2000) 47–50. [7] Y. Arita, S. Kihara, N. Ouchi, K. Maeda, H. Kuriyama, Y. Okamoto, M. Kumada, K. Hotta, M. Nishida, M. Takahashi, T.
P.W. Peake et al. / Biochemical and Biophysical Research Communications 367 (2008) 560–565
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Nakamura, I. Shimomura, M. Muraguchi, Y. Ohmoto, T. Funahashi, Y. Matsuzawa, Adipocyte-derived plasma protein adiponectin acts as a platelet-derived growth factor-BB-binding protein and regulates growth factor-induced common postreceptor signal in vascular smooth muscle cell, Circulation 105 (2002) 2893–2898. P.W. Peake, Y. Shen, L.V. Campbell, J.A. Charlesworth, Human adiponectin binds to bacterial lipopolysaccharide, Biochem. Biophys. Res. Commun. 341 (2006) 108–115. Y. Wang, K.S. Lam, J.Y. Xu, G. Lu, L.Y. Xu, G.J. Cooper, A. Xu, Adiponectin inhibits cell proliferation by interacting with several growth factors in an oligomerization dependent manner, J. Biol. Chem. 280 (2005) 18341–18347. Y. Takemura, N. Ouchi, R. Shibata, T. Aprahamian, M.T. Kirber, R.S. Summer, S. Kihara, K. Walsh, Adiponectin modulates inflammatory reactions via calreticulin receptor-dependent clearance of early apoptotic bodies, J. Clin. Invest. 117 (2007) 375–386. P.W. Peake, J.T. Hughes, Y. Shen, J.A. Charlesworth, Glycosylation of human adiponectin affects its conformation and stability, J. Mol. Endocrinol. 39 (2007) 45–52. F.D. McGrath, M.C. Brouwer, G.J. Arlaud, M.R. Daha, C.E. Hack, A. Roos, Evidence that complement protein C1q interacts with Creactive protein through its globular head region, J. Immunol. 176 (2006) 2950–2957. P.W. Peake, A.D. Kriketos, L.V. Campbell, Y. Shen, J.A. Charlesworth, The metabolism of isoforms of human adiponectin: studies in human subjects and in experimental animals, Eur. J. Endocrinol. 153 (2005) 409–417. S.S. Bohlson, D.A. Fraser, A.J. Tenner, Complement proteins C1q and MBL are pattern recognition molecules that signal immediate and long-term protective immune functions, Mol. Immunol. 44 (2007) 33–43. U. Kishore, C. Gaboriaud, P. Waters, A.K. Shrive, T.J. Greenhough, K.B. Reid, R.B. Sim, G.J. Arlaud, C1q and tumor necrosis factor superfamily: modularity and versatility, Trends Immunol. 25 (2004) 551–561. L.T. Roumenina, A.A. Kantardjiev, B.P. Atanasov, P. Waters, M. Gadjeva, K.B. Reid, A. Mantovani, U. Kishore, M.S. Kojouharova, Role of Ca2+ in the electrostatic stability and the functional activity of the globular domain of human C1q, Biochemistry 44 (2005) 14097– 14109. R. Ghai, P. Waters, L.T. Roumenina, M. Gadjeva, M.S. Kojouharova, K.B. Reid, R.B. Sim, U. Kishore, C1q and its growing family, Immunobiology 212 (2007) 253–266.
565
[18] Y. Wang, K.S. Lam, L. Chan, K.W. Chan, J.B. Lam, M.C. Lam, R.C. Hoo, W.W. Mak, G.J. Cooper, A. Xu, Post-translational modifications of the four conserved lysine residues within the collagenous domain of adiponectin are required for the formation of its high molecular weight oligomeric complex, J. Biol. Chem. 281 (2006) 16391–16400. [19] A.A. Richards, T. Stephens, H.K. Charlton, A. Jones, G.A. Macdonald, J.B. Prins, J.P. Whitehead, Adiponectin multimerization is dependent on conserved lysines in the collagenous domain: evidence for regulation of multimerization by alterations in posttranslational modifications, Mol. Endocrinol. 20 (2006) 1673–1687. [20] Y. Wang, L.Y. Xu, K.S. Lam, G. Lu, G.J. Cooper, A. Xu, Proteomic characterization of human serum proteins associated with the fatderived hormone adiponectin, Proteomics 6 (2006) 3862–3870. [21] A. Sjoberg, P. Onnerfjord, M. Morgelin, D. Heinegard, A.M. Blom, The extracellular matrix and inflammation: fibromodulin activates the classical pathway of complement by directly binding C1q, J. Biol. Chem. 280 (2005) 32301–32308. [22] U. Kishore, S.K. Gupta, M.V. Perdikoulis, M.S. Kojouharova, B.C. Urban, K.B. Reid, Modular organization of the carboxyl-terminal, globular head region of human C1q A, B, and C chains, J. Immunol. 171 (2003) 812–820. [23] M.S. Kojouharova, M.G. Gadjeva, I.G. Tsacheva, A. Zlatarova, L.T. Roumenina, M.I. Tchorbadjieva, B.P. Atanasov, P. Waters, B.C. Urban, R.B. Sim, K.B. Reid, U. Kishore, Mutational analyses of the recombinant globular regions of human C1q A, B, and C chains suggest an essential role for arginine and histidine residues in the C1q–IgG interaction, J. Immunol. 172 (2004) 4351–4358. [24] L.A. Trouw, A.M. Blom, P. Gasque, Role of complement and complement regulators in the removal of apoptotic cells, Mol. Immunol. 45 (2008) 1199–1207. [25] K. Duus, R.T. Pagh, U. Holmskov, P. Hojrup, S. Skov, G. Houen, Interaction of calreticulin with CD40 ligand, TRAIL and Fas ligand, Scand. J. Immunol. 66 (2007) 501–507. [26] R.W. Vandivier, C.A. Ogden, V.A. Fadok, P.R. Hoffmann, K.K. Brown, M. Botto, M.J. Walport, J.H. Fisher, P.M. Henson, K.E. Greene, Role of surfactant proteins A, D, and C1q in the clearance of apoptotic cells in vivo and in vitro: calreticulin and CD91 as a common collectin receptor complex, J. Immunol. 169 (2002) 3978–3986. [27] Y.S. Kang, Y. Do, H.K. Lee, S.H. Park, C. Cheong, R.M. Lynch, J.M. Loeffler, R.M. Steinman, C.G. Park, A dominant complement fixation pathway for pneumococcal polysaccharides initiated by SIGN-R1 interacting with C1q, Cell 125 (2006) 47–58.