Interaction of the C1 complex of Complement with sulfated polysaccharide and DNA probed by single molecule fluorescence microscopy Be ´ range ` re Tissot 1 ,Re ´ gis Daniel 1 and Christophe Place 2 1 Laboratoire Analyse et Environnement, Universite ´ d’Evry, France; 2 Laboratoire de Physique, Ecole Normale Supe ´ rieure de Lyon, France The complex C1 triggers the activation of the Complement classical pathway through the recognition and binding of antigen–antibody complex by its subunit C1q. The globular region of C1q is responsible for C1 binding to the immune complex. C1q can also bind nonimmune molecules such as DNA and sulfated polysaccharides, leading either to the activation or inhibition of Complement. The binding site of these nonimmune ligands is debated in the literature, and it has been proposed to be located either in the globular region or in the collagen-like region of C1q, or in both. Using single molecule fluorescence microscopy and DNA molecular combing as reporters of interactions, we have probed the C1q binding properties of T4 DNA and of fucoidan, an algal sulfated fucose-based polysaccharide endowed with potent anticomplementary activity. We have been able to visualize the binding of C1q as well as of C1 and of the isolated collagen-like region to individual DNA strands, indicating that the collagen-like region is the main binding site of DNA. From binding assays with C1r, one of the protease compo- nents of C1, we concluded that the DNA binding site on the collagen-like region is located within the stalk part. Com- petition experiments between fucoidan and DNA for the binding of C1q showed that fucoidan binds also to the col- lagen-like region part of C1q. Unlike DNA, the binding of fucoidan to collagen-like region involves interactions with the hinge region that accommodate the catalytic tetramer C1r 2 –C1s 2 of C1. This binding property of fucoidan to C1q provides a mechanistic basis for the anticomplementary activity of the sulfated polysaccharide. Keywords: fucoidan; C1q; complement system; single mole- cule fluorescence microscopy. Studies on the interactions between carbohydrates and proteins represent a major and challenging topic in glyco- biology, as it is now recognized that many crucial life processes are dependent on their specific molecular recog- nition. Carbohydrate–protein interactions mediate funda- mental biological mechanisms, encompassing growth control, apoptosis, cell differentiation and proliferation, as well as physiopathologic disorders like tumoral metastasis, autoimmune diseases and inflammation [1,2]. However the mechanisms of these interactions involving carbohydrates are still poorly understood, particularly with regard to the molecular basis of the strength and specificity of binding to targeted proteins [3]. Difficulties mainly arise from the high structural diversity and from the complex dynamic proper- ties of polysaccharides [4]. A new approach based on single- molecule detection is currently arousing great interest in biology as it allows the direct observation and manipulation of biomolecules [5,6]. This approach has already been successfully applied to the study of the interaction of DNA and proteins [7,8]. Comparatively few data have been reported on the study of carbohydrates and their interaction by this approach [9,10], probably because of the difficulties in manipulation of such structurally heterogeneous and polydisperse biopolymers at the single-molecule level. Most of the studies provide topographic images by atomic force microscopy of polysaccharide molecules adsorbed on a surface [6,11]. Nevertheless, we think that new insights into the carbohydrate–protein interactions should be obtained by studying them at a single-molecule level in terms of the protein partner. We have applied this strategy to the study of the bioactive polysaccharide fucoidan, one of the most potent inhibitors of the human Complement system. Fucoidan is a sulfated polysaccharide extracted from brown algae and present- ing a structural organization based on an [fi4)-a- L -Fucp-(1fi3)-a- L -Fucp-(1fi4)-a- L -Fucp-(1fi3)-a- L -Fucp (1fi] backbone [12,13]. It is assumed that its biological properties are related to its capacity to achieve specific interactions with targeted proteins. We have recently shown that fucoidan inhibits the first steps of activation of the Complement cascade [14]. In addition, affinity electrophor- esis experiments indicated interaction between fucoidan and the Complement protein C1q, and this interaction could result in the observed inhibition [14]. The protein C1q, a subunit of the C1 complex, is involved in the recognition Correspondence to R. Daniel, Laboratoire Analyse et Environnement, UMR 8587 CNRS, Universite ´ d’Evry-Val-d’Essonne, Bd. Franc¸ ois Mitterrand, 91025 Evry cedex, France. Fax: + 33 1 69477655, Tel.: + 33 1 69477641, E-mail: regis.daniel@chimie.univ-evry.fr Abbreviations: CLR, collagen-like region; EDC, 1-ethyl-3- (3-dimethylaminopropyl) carbodiimide hydrochloride; GR, globular region; PMMA, polymethylmethacrylate; sulfo-NHS, sulfo-N-hydroxysuccinimide. (Received 23 July 2003, revised 22 September 2003, accepted 6 October 2003) Eur. J. Biochem. 270, 4714–4720 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03870.x and the binding of antigen–antibody complexes that triggers the activation of the Complement [15]. C1q (460 kDa) made of three polypeptide chains (A, B and C) exhibits unique structural features (Fig. 1). It consists of a C-terminus presenting six globular head groups connected through a hinge region to a long (approximately 11 nm) triple helical collagen-like stalk that ends at the N-terminus [16]. Because of this structural organization, C1q is often pictured as a bunch of six flowers. The interaction of C1q with immune complexes takes place at the globular region [17,18], but C1q is known to also bind through its collagen-like region (CLR) several nonimmune molecules [19], with conse- quences which remain unclear. Actually the binding of the C-reactive protein [20], of the serum amyloid protein [21] and of DNA [22] to C1q leads to an activation of Complement. On the other hand, the binding of sulfated glycosaminoglycans [23–25], of proteoglycan dermatan sulfate decorin [26], and of chondroitin 4-sulfate (i.e. the C1q inhibitor) results in the inhibition of the classical pathway [27]. Our goal in this study is to ascertain the binding of fucoidan to C1q and to determine the site of interaction on the protein. For this purpose we took advantage of the binding properties of C1q toward DNA and of an emerging technique allowing the molecular combing of DNA strands and its observation by single-molecule spectroscopy. Dou- ble- and single-stranded DNA has been demonstrated to bind preferentially to the collagen-like region of C1q under physiological saline conditions [22,28,29]. A cationic peptide sequence on the A chain of C1q has been identified as the major binding site of DNA [30]. We have analyzed herein the binding of the human of the C1 complex and of its subunit C1q to T4 DNA by molecular combing which results in a large array of DNA strands individually observed by fluorescence microscopy [8]. This approach allowed us to implement an analytical tool useful to investigate the binding of fucoidan to C1q through compe- tition experiments between DNA and the polysaccharide. We have addressed the question of whether a C1q inhibitor (fucoidan) and a C1q activator (DNA) are able to bind to the same region of the protein by using not only native C1q but also the C1q isolated domains CLR and the globular region (GR). Materials and methods Buffers The following buffers were used: 250 m M Bis/Tris, pH 6.47; 0.1 M Mes buffer, pH 6, containing 0.5 M NaCl; and 1 M sodium hydrogen carbonate (NaHCO 3 ) buffer, pH 8.4. All buffers were prepared with ultrapure water (milliQ, Milli- pore). Reagents and proteins The Complement proteins C1r, C1q and C1 as well as the depleted sera and the specific antibodies were obtained from VWR (Fontenay-sous-Bois, France). C1q designed as purified C1q in this study, and the derived collagen-like region CLR and globular heads region GR were a generous gift from G.J. Arlaud (IBS, Grenoble, France). The CLR and GR were prepared as previously described [28]. Double- stranded DNA from salmon testes (type III), T4 DNA and 2-mercaptoethanol were purchased from Sigma (Saint- Quentin Fallavier, France). 1-Ethyl-3-(3-dimethylamino- propyl) carbodiimide hydrochloride (EDC) and sulfo-NHS (sulfo-N-hydroxysuccinimide) were purchased from Pierce (Rockford, IL, USA). The fluorescent intercalating agent YOYO-1 (dimer yellow oxozalone) and the Alexa 488- fluorescent beads (27 nm) were obtained from Molecular Probes (Eugene, OR, USA). Fucoidan from the brown algae Ascophyllum nodosum was prepared as previously described [31]. The fucoidan fraction used for the study herein was of low molecular weight (8000 gÆmol )1 as determined by high performance size exclusion chromato- graphy using heparin standards [32]) and of high sulfate content (30% w/w), and was endowed with a high anti- complementary activity as we have previously reported [14]. Fig. 1. Schematic representation of the structural organization of human C1. (A) Model of the C1 complex showing the catalytic tetramer C1r 2 – C1s 2 interacting with C1q (adapted from [37]). (B) Representation of the association between the A, B and C chains constituting the six ABC heterotrimeric triple helices of C1q. The cationic sequence 14–26 of the A chain, as well as the sequence responsible for the kink of the colla- gen-like region are shown. Ó FEBS 2003 Complement C1 complex interactions (Eur. J. Biochem. 270) 4715 Surface treatment Glass surfaces were rendered hydrophobic by coating with polymethylmethacrylate (PMMA, M r 8000 gÆmol )1 ). A droplet of PMMA in chlorobenzene (13% w/w) is put down onto a clean glass cover-slide and spread with a spin coater of in-house design, at 2500 r.p.m. for 3 min. Surfaces are then baked at 165 °C for 40 min and stored at room temperature in a dust-free environment. DNA preparation T4 DNA (160 kb) was labeled as follows: 7.2 lLof1.35n M T4 DNA were incubated with 10 l M YOYO-1 (Molecular Probes) in ultrapure water. This respresents a ratio of 1 molecule of dye to 30 bases of DNA, and 150 lLBis/Tris pH 6.47, completed with 1.5 mL of ultrapure water during at least 1 h at room temperature. Fluorescent beads preparation Alexa 488-conjugated beads (Molecular Probes) were activated following the manufacturers’ instructions. Briefly, 50 lLofbeads(3.25l M in ultrapure water) were mixed with 50 lL of sulfo-NHS (0.5 M in ultrapure water) and 10 lLofEDC(0.2 M in ultrapure water) in 100 lLof0.1 M Mes buffer and completed with 400 lL of ultrapure water. After 60 min incubation at room temperature under gentle agitation, reaction was stopped by addition of 3 lLof 2-mercaptoethanol. Elimination of the excess of reactants was performed on P6 Biospin Columns (Bio-Rad). The concentration of used beads solutions in 0.1 M Mes buffer ranged from 20 to 30 n M , as evaluated by spectrophoto- metric determination at 505 nm. Protein labeling Proteins C1q, C1r and the CLR fragments of C1q were labeled with fluorescent beads. The concentration of beads was adjusted according to the type and the concentration of protein in order to maximize the rate of the labeling reaction without generating cross-linking of the beads. Twenty microliters of commercial C1q (2.46 l M ) were added to 0.5 lL of beads solution (25 n M )andto2.5lLNaHCO 3 buffer completed with 25 lL ultrapure water. After 45 min incubation at 20 °C under gentle agitation, reaction was stopped by addition of 0.6 lLNH 2 OH (3 M in ultrapure water). Purified C1q (2.8 l M ;17.5 lL) were added to 0.5 lL of beads solution (25 n M )andto25lLNaHCO 3 buffer completed with 100 lL ultrapure water. After 45 min at 20 °C under gentle agitation, reaction was stopped by addition of 2.4 lLNH 2 OH (3 M in ultrapure water). Six microliters of CLR (7.8 l M ) were added to 0.5 lLofbeads solution (25 n M )andto12.5lLofNaHCO 3 buffer completed with 100 lL ultrapure water. After 45 min at 20 °C under gentle agitation, reaction was stopped by addition of 2.4 lLofNH 2 OH (3 M in ultrapure water). Twenty microlitres of C1 (0.27 l M )wereaddedto0.5lL of beads solution (20 n M )andto2.5lLNaHCO 3 buffer completed with 25 lL ultrapure water. After 45 min at 20 °C under gentle agitation, reaction was stopped with 1 lLofNH 2 OH (3 M in ultrapure water). Two microliters of C1r (10.5 l M )wereaddedto1lL of bead solution (25 n M )andto4lLNaHCO 3 buffer completed with 40 lL ultrapure water. After 45 min at 20 °C under gentle agitation, reaction was stopped by addition of 2 lL NH 2 OH (3 M in ultrapure water). Fluid phase incubations of DNA with complement proteins Fluorescent T4 DNA was incubated with the labeled proteins, either in the presence of fucoidan or not, for 45–60 min at room temperature under gentle agitation in the following conditions: 150 lLof6.5p M T4 DNA were mixedwith1lL of labeled C1q prepared as above, and according to case with 5 lLof45l M fucoidan; when the purified C1q was used, 400 lLof6.5p M T4 DNA were mixedwith10 lL of labeled purified C1q. One hundred and fifty microliters of 6.5 p M T4 DNA were mixed with 5 lLof labeled CLR or 2.5 lL of labeled C1, and according to case with 1–2 lLof45 l M fucoidan. Then, the treated DNA was combed as described below. For the study of the interaction between C1r and C1q bound to combed T4 DNA, 150 lL of 6.5 p M T4 DNA were preincubated with 1 lLof2l M unlabeled C1q for 45 min, then 6 lL of labeled C1r were added, and according to case with 9 lLof45l M fucoidan. Molecular combing The combing process consists in the stretching of the DNA by the passage of an air/water meniscus [33,34]. A droplet of 6.5 p M T4 DNA in Bis/Tris buffer, pH 6.47, is deposed on a hydrophobic surface, incubated for 2 min and then removed. This procedure is sufficient to stretch the DNA strands when DNA is brought in small droplet. The interaction between DNA and the hydrophobic surface is very strong, so that DNA can be considered as grafted onto the surface. Fluorescent microscopy Samples were observed using an inverted microscope (Leica DM IRBE) by epifluorescence. An X100 infinity-corrected 1.4 NA oil objectives (Leica) was used and a cooled CCD camera (C4880 Hamamatsu, and Ixon-Andor) was moun- ted on the microscope. For fluorescence observations, a mercury lamp was used in combination with a filters set for fluorescein (Leica I3). The images were acquired using the HIPIC software (Hamamatsu) and IXON software (Andor), with an exposition time ranging from 100 ms to 1 s. Results Interaction between DNA and C1q DNA is described as an activator of human complement system [30]. The DNA-dependent activation of Comple- ment may result from the formation of a complex between DNA and the first Complement protein C1q as proposed in the literature [20,22,28]. The possibility to visualize individ- ual DNA molecules prompted us first to investigate the binding properties between C1q and DNA through the observation of their complex. 4716 B. Tissot et al. (Eur. J. Biochem. 270) Ó FEBS 2003 T4 DNA (160 kbps) labeled with the fluorescent inter- calating agent YOYO-1 was combed on a PMMA surface. Fluorescent strands were observed corresponding to indi- vidual linear DNA chains stretched on the surface as previously described (Fig. 2A) [8]. Fluorescent T4 DNA was incubated in solution with labeled C1q. For that purpose, C1q was covalently attached through its amino groups to fluorescent beads (27 nm diameter). C1q is estimated to have an overall size of 35 nm based on literature data [15]. Given their similar sizes, we can reasonably expect that one or two molecules of C1q are bound per fluorescent bead. After incubation with such labeled C1q (C1q*), fluorescent T4 DNA was combed as described above. The fluorescent individual DNA molecules then observed were clearly decorated with a succession of fluorescent spots (Fig. 2B). The fluorescence of these spots is easily distinguishable from the YOYO-1-induced fluores- cence of the combed DNA, by its color and by its strong intensity corresponding to the high density of the Alexa fluor contained in the beads. These fluorescent spots were then due to the presence of the beads along the DNA strands. When the beads were noncoupled to C1q, no such binding was observed on DNA strands (data not shown). These results indicate that C1q mediates the attachment of the spheres to DNA, hence evidencing interaction between C1q and DNA. In order to identify the preferential DNA binding region on C1q, we have carried out experiments with the separated domains of C1q, i.e. the collagen-like region CLR and the globular region GR. CLR and GR were prepared by enzymatic digestion by pepsin and collagenase, respectively, as previously described [28]. The C-terminal domain of the CLR, named the hinge region and joining the CLR to the GR, is resistant to both proteases so that the conforma- tional organization of CLR is conserved, whereas the GR was obtained as individual globular heads [35]. In a first set of experiments, CLR and GR were studied for their capacity to compete with purified C1q* for the binding to DNA. When incubation of T4 DNA and C1q* was performed with various amounts of CLR, the analysis of the resulting combed DNA showed that the binding of C1q* to DNA strands started to decrease for a C1q/CLR ratio of 1 : 10 and was totally suppressed from a 1 : 50 molar ratio. On the other hand, a higher amount of GR was required to observe a similar inhibition as a C1q*/GR molar ratio of 1 : 3000 was at least necessary (data not shown). Such a large difference in the amount required to efficiently compete for the binding of C1q* indicates that the collagen- like region of C1q contains the preferential site for the interaction and the binding to DNA. In order to confirm this result, we checked for the binding of the CLR to DNA using CLR preparation labeled with the fluorescein-conjugated beads (CLR*). CLR* was incu- bated with T4 DNA in fluid phase, after which the DNA was combed on a PMMA surface. The analysis of the images showed that CLR* was colocalized like C1q* with the combed DNA (Fig. 2C), proving the binding of CLR to DNA. The C1 complex, which triggers the classical pathway of Complement, comprises the subunit C1q and the two serine proteases C1r and C1s associated into a C1r 2 –C1s 2 tetramer. Several lines of evidence in literature indicate that the binding site of the tetramer on C1q is located within the collagen-like region of C1q [36]. As our results show that this region also contains the binding site of DNA, we wondered whether the association of the tetramer C1r 2 – C1s 2 to C1q could interfere with the binding of C1q to DNA. We studied the binding to DNA of the C1 complex labeled with the fluorescent beads (C1*). We observed that C1* and the individual DNA strands were colocalized (Fig. 2D), indicating the binding of C1 to DNA. Hence the presence of the tetramer bound to the collagen-like region of C1q does not impede the binding of C1q to DNA. Conversely, in a subsequent experiment, we checked the ability of C1r to associate on DNA-bound C1q. For that purpose, T4 DNA was preincubated with nonlabeled C1q, in order to form a DNA–C1q complex. The mixture was then incubated with C1r labeled with fluorescent beads (C1r*), and finally DNA was combed. We observed that C1r* spots were aligned along the DNA strands (Fig. 2E). Thus C1r binds to the DNA–C1q assembly, whereas C1r does not bind to DNA strands in the absence of C1q (data not shown). Therefore the binding of C1r is a direct consequence of the presence of C1q on DNA strands. Fig. 2. Molecular combing of the T4 DNA on a PMMA surface after incubation with fluores- cent Complement proteins and fucoidan. (A) Individual strands of fluorescent 160 kbps T4 DNA combed on the PMMA surface. (B–D) Molecular combing of the T4 DNA after incubation with fluorescent (Alexa 488-fluor- escent beads) C1q, CLR and C1 proteins, respectively. (E) Molecular combing of the T4 DNA after incubation, first with nonlabeled C1q and then with fluorescent C1r. (F) Molecular combing of the T4 DNA after incubation with fluorescent C1q in presence of fucoidan. Ó FEBS 2003 Complement C1 complex interactions (Eur. J. Biochem. 270) 4717 Altogether these results indicate that the collagen-like region of C1q contains distinct sites for the binding of DNA and for the binding of the catalytic tetramer C1r 2 –C1s 2 . Interactions between fucoidan and C1q We have previously reported that the sulfated polysaccha- ride fucoidan interacts with C1q. In order to ascertain this binding property of fucoidan, we performed competition experiments between fucoidan and DNA for the binding to C1q. Fluorescent T4 DNA was incubated with C1q* and fucoidan in the C1q/fucoidan molar ratio 1 : 100 (i.e. the amount which leads to 30% inhibition of the hemolytic activity of C1q as we have previously reported [14]). After combing, the individual fluorescent DNA strands clearly appeared without any decoration by C1q* (Fig. 2F). This result shows that fucoidan is able to compete with DNA for the binding to C1q, suggesting that interactions between fucoidan and C1q probably occur through the collagen-like region. This hypothesis was confirmed by performing the same competition experiment using labeled-CLR (CLR*). Incubation of T4 DNA with CLR* in the presence of fucoidan (CLR/fucoidan molar ratio of 1 : 20 and 1 : 40) leads to an inhibition of the binding of CLR* to DNA from the molar ratio 1 : 20. Altogether these results confirm that fucoidan interacts with C1q through the same binding region than DNA, i.e. the collagen-like region that includes in our CLR preparation the stalk region and partially the hinge region. At this stage, it is worthwhile to note that, when C1 instead of C1q was used in this competition experiment with fucoidan, a lower inhibition of the protein binding to DNA was obtained (for the same C1/fucoidan molar ratio 1 : 100), as colocalization of C1* was still observed with some DNA strand (Fig. 3). This lower inhibition may be due to the presence of the catalytic tetramer C1r 2 –C1s 2 in the C1 complex that hinders the interaction of the C1q subunit with the polysaccharide. According to this hypo- thesis fucoidan should then interact with the hinge region containing the binding site of the tetramer C1r 2 –C1s 2 ,in addition to the stalk region of CLR. In order to check this hypothesis, we tested the effect of fucoidan on the binding of C1r to DNA-bound C1q. For this purpose, T4 DNA and nonlabeled C1q were firstly preincubated, before the addition of C1r* and fucoidan (molar ratio C1q/fucoidan 1 : 100). The resulting molecular combing appeared as in Fig. 2F, exhibiting the absence of C1r* spots on the T4 strands. We have seen above that C1r is able to associate to DNA-bound C1q. Furthermore, we have previously repor- ted an affinity capillary electrophoresis study evidencing no interaction between either C1r or C1s and fucoidan [14]. Hence this result proved that the binding of C1r to C1q is inhibited by fucoidan, consistent with the interaction of the polysaccharide with the hinge region of the collagen-like region of C1q. Discussion C1q can bind several polyanionic molecules like sulfated polysaccharides and also DNA, but with the opposite effects of either inhibition or activation, respectively. Using single-molecule observation of immobilized DNA strands, we have been able to visualize the binding of C1q to individual DNA strands. This single molecule approach appeared as a valuable tool with which to investigate the binding properties of fucoidan, a sulfated fucose-based polysaccharide known as one of the most potent inhibitor of IgG-dependent activation of Complement [14]. C1q binding to DNA was deduced in literature from data based either on C1q–DNA precipitation experiments [22] or on solid-phase assays with immobilized C1q [28,29]. In the present study, we obtained the direct evidence of the DNA– C1q binding through the observation of the colocalization of C1q and DNA strands. It is worthwhile noting that this interaction was performed in the fluid phase. In these conditions, the resulting C1q-DNA complex was resistant to the stretching and combing of DNA on PMMA surfaces, indicating the strength of the interaction. A continuous succession of fluorescent C1q could be observed on some DNA strands, corresponding to a high density of binding. Although the optical resolution does not allow the deter- mination of this density, the obtained images are consistent with previous estimates of one C1q per 34 nm of double- stranded DNA for the highest density [22]. C1q is usually described as constituted of two main domains, the N-terminal collagen-like region CLR, and the C-terminal globular region GR. Divergent data have been reported in the literature concerning the binding site of DNA on C1q. DNA has been proposed to bind to either the GR or the CLR, or to both of the two domains, according the method used. Here, we have individually used each of these domains in the single molecule approach to unambiguously identify the CLR as the main binding site of DNA. Indeed, we have shown that CLR binds to DNA as well as C1q does, and that, compared with GR, CLR competes much more efficiently with C1q for binding to DNA. The dissociation of the globular region of C1q into individual structures could decrease the affinity of the GR, but to an extent that could not lead to the observed difference between CLR and GR in these competition experiments. The low competition effect observed only when a very large excess of GR is used could be due to non-specific interactions or to interaction with the residual GR tail present in the GR preparation. The collagen-like region can be divided into an N-terminal domain organized into a triple-helical stalk, which diverges into six arms constituting the C-terminal Fig. 3. Molecular combing of the T4 DNA on a PMMA surface after incubation with fluorescent C1 complex and fucoidan. 4718 B. Tissot et al. (Eur. J. Biochem. 270) Ó FEBS 2003 domain of the CLR; this has been named the hinge region [15]. This hinge region contains the binding site for the catalytic tetramer C1r 2 –C1s 2 , essential for the C1 activity [35]. Our results show that C1q as well as C1 bind to DNA, and that C1r can bind to C1q bound to DNA. Thus DNA binding site on CLR does not overlap with the binding site of the tetramer in the hinge region. It is likely that DNA binds to the stalk domain of CLR, consistent with previous findings showing that a synthetic peptide of the N-terminal portion of the A chain (residues 14–26) binds to DNA and inhibits its binding to C1q [20,30]. During the course of our studies of the anticomplemen- tary activity of fucoidan, we have previously shown by affinity electrophoresis that this sulfated polysaccharide binds also to C1q [14]. The results obtained here clearly showed its ability to inhibit the binding of C1q, as well as that of CLR, to DNA. These data indicate that, like DNA, fucoidan binds to C1q through the collagen-like region. The C1q A chain that appeared to be essential for the binding of nonimmune substance contained a cationic region within residues 14–26 of stalk [20]. This positively charged sequence contained five basic proximal residues, arginine and lysine, that are assumed to be involved in the binding of polyanions like DNA and fucoidan (Fig. 1B). Consistent with these data, we have previously shown that fucoidan protects the lysine residues of C1q from chemical modification by specific reagent [14]. However fucoidan exhibits a major difference with DNA in that the polysaccharide is able to block the association of C1r to C1q. This result is in agreement with our previous finding where the polysaccha- ride was shown by ELISA to inhibit the reconstitution of C1 from C1r, C1s, and C1q [14]. Thus, unlike DNA, fucoidan also interacts with the Ôarms domainÕ of CLR, which contains the binding site of the tetramer C1r 2 –C1s 2 .Ithas been reported that basic residues lysine and arginine in the hinge region are involved in the assembly of C1 through specific interaction with acidic residues of C1r [35]. Furthermore, the structural model of C1q shows a cluster of basic residues that are located in the hinge region at the junction between the stalk and the arms. We assumed that fucoidan interacts with these positively charged residues in the hinge region, leading to the observed blockage of the C1 assembly. As a consequence, the inhibiting activity of fucoidan on the classical pathway activation should result from this binding property to the hinge region, hampering the activation of the two proteases C1r and C1s. This mechanism is probably related to the inhibiting property of endogenous C1q inhibitors of Complement, like the chon- droitin 4-sulfate proteoglycan, which is secreted by the human B lymphocytes. This glycosaminoglycan has been proposed as a potential physiologic C1q inhibitor, through the inhibition of the C1q– (C1r 2 –C1s 2 ) assembly [37]. Other C1q binding substances that are not C1q inhibi- tors, like DNA, C-reactive protein [20], and amyloid protein, do not bind to the hinge region but rather to the stalk domain of the CLR [38]. It has also been proposed that the collagen-like stalk is involved in the binding of C1q to different cell types and to liposomes [39]. Strikingly, this binding leads to the activation of Complement, as we observed for DNA (data not shown). The mechanism of this activation, independent of the recognition of the immune complex by the globular heads, remains unclear and is debated in the literature [40]. Recently, it has been reported from the structural model of C1r that the activation of the C1 complex could result from mechanical constraints upon C1q binding, which affect the flexible hinge region [41]. Further investigations are required to determine whether the binding to the stalk region also results in such mechanical stresses that could be transmitted to the hinge region. Acknowledgements We thank J. Ratiskol and C. Sinquin for the extraction and the preparation of the fucoidan fraction and for their experimental advices. We are grateful to Professor G. J. Arlaud for his generous gift of purified C1q and of GR and CLR preparations. This work was supported by CNRS and the county Pays de La Loire, as well as the program Physique et Chimie du Vivant, from CNRS and the Ministe ` re de la Recherche, France. References 1. Williams, S.J. & Davies, G.J. (2001) Protein–carbohydrate inter- actions: learning from nature. Trends Biotechnol. 19, 356–362. 2. Lee, Y.C. & Lee, R.T. (1995) Carbohydrate–protein interactions: basis of glycobiology. Acc. Chem. Res. 28, 321–327. 3. Spillmann, D. & Lindahl, U. (1994) Glycosaminoglycan–protein interactions: a question of specificity. Curr. Opin. Struct. Biol. 4, 677–682. 4. Imberty, A. & Perez, S. (2000) Structure, conformation, and dynamics of bioactive oligosaccharides: theoretical approaches and experimental validations. Chem. Rev. 100, 4567–4588. 5. Mehta, A.D., Rief, M., Spudlich, J.A., Smith, D.A. & Simmons, R.M. (1999) Single-molecule biomechanics with optical methods. Science 238, 1689–1695. 6. Ludwig, M., Rief, M., Schmidt, L., Li, H., Oesterhelt, F. & Gautel, M. (1999) AFM, a tool for single-molecule experiments. Appl.Phys.Mater.Sci.Process.68, 173–176. 7. Ishijima, A. & Yanagida, T. (2001) Single molecule nano- bioscience. Trends Biochem. Sci. 26, 438–444. 8. Gueroui, Z., Place, C., Freyssingeas, E. & Berge, B. (2002) Observation by fluorescence microscopy of transcription on single combed DNA. Proc. Natl Acad. Sci. USA 99, 6005–6010. 9. Brant, D.A. (1999) Novel approaches to the analysis of polysac- charide structures. Curr. Opin. Struct. Biol. 9, 556–562. 10. Marzalek, P., Li, H. & Fernandez, J. (2001) Fingerprinting poly- saccharides with single-molecule atomic force microscopy. Nat. Biotechnol. 19, 258–262. 11. Camesano, T. & Wilkinson, K. (2001) Single molecule study of xanthan conformation using atomic force microscopy. Biomacromolecules 2, 1184–1191. 12. Daniel,R.,Berteau,O.,Jozefonvicz,J.&Goasdoue,N.(1999) Degradation of algal (Ascophyllum nodosum) fucoidan by an enzymatic activity contained in digestive glands of the marine mollusk Pecten maximus. Carbohydr. Res. 322, 291–297. 13. Chevolot, L., Mulloy, B., Ratiskol, J., Foucault, A. & Colliec- Jouault, S. (2001) A disaccharide repeat unit is the major structure in fucoidans from two species of brown algae. Carbohydr. Res. 330, 529–535. 14. Tissot, B., Montdargent, B., Chevolot, L., Varenne, A., Descroix, S., Gareil, P. & Daniel, R. (2003) Interaction of fucoidan with the complement proteins of the classical pathway. Biochem. Biophys. Acta 1651, 5–16. 15. Cooper, N.R. (1985) The classical complement pathway: activa- tion and regulation of the first complement component. Adv. Immunol. 37, 151–216. Ó FEBS 2003 Complement C1 complex interactions (Eur. J. Biochem. 270) 4719 16. Kishore, U. & Reid, K.B. (2000) C1q: structure, function, and receptors. Immunopharmacology. 49, 159–170. 17.Marques,G.,Anton,L.C.,Barrio,E.,Sanchez,A.,Ruiz,S., Gavilanes, F. & Vivanco, F. (1993) Arginine residues of the globular regions of human C1q involved in the interaction with Immunoglobulin G. J. Biol. Chem. 268, 10393–10402. 18. Kaul, M. & Loos, M. (1997) Dissection of C1q capability of interacting with IgG. J. Biol. Chem. 272, 33234–33244. 19. Gewurz, H., Ying, S.C., Jiang, H. & Lint, T.F. (1993) Non immune activators of the classical complement pathway. Behring Institute Mitt. 93, 138–147. 20. Jiang, H., Robey, F. & Gewurz, H. (1992) Localization of sites through which C-reactive protein binds and activates complement to residues 14–26 and 76–92 of the human C1q A chain. J. Exp. Med. 175, 1373–1379. 21. Ying, S Y., Gewurz, A.T., Jiang, H. & Gewurz, H. (1993) Human serum amyloid P component oligomers bind and activate the classical complement pathway via residues 14–26 and 76–92 of the A chain collagen-like region of C1q. J. Immunol. 150, 169–176. 22. Schravendijk, M.R. & Dwek, R.A. (1982) Interaction of C1q with DNA. Mol. Immunol. 19, 1179–1187. 23. Calabrese, G., Recondo, M., Fernandez de Recondo, M. & Recondo, E. (1997) The first component of the human comple- ment system recognizes the active fraction of heparin. Cell. Mol. Biol. 43, 237–242. 24. Weiler, J. & Linhardt, R. (1989) Comparison of the activity of polyanions and polycations on the classical and the alternative pathways of complement. Immunopharmacology 17, 65–72. 25. Sahu, A. & Pangburn, M.K. (1993) Identification of multiple sites of interaction between heparin and the complement system. Mol. Immunol. 30, 679–684. 26. Krumdieck, R., Hook, M., Rosenberg, L. & Volanakis, J. (1992) The proteoglycan decorin binds C1q and inhibits the activity of the C1 complex. J. Immunol. 149, 3695–3701. 27. Silvestri, L., Baker, J., Roden, L. & Stroud, R. (1981) The C1q inhibitor in serum is a chondroitin 4-sulfate proteoglycan. J. Biol. Chem. 256, 7383–7387. 28. Rosenberg, A.M., Prokopchuk, P.A. & Lee, J.S. (1988) The binding of native DNA to the collagen-like segment of C1q. J. Rheumatol. 15, 1091–1096. 29. Uwatoko, S. & Mannik, M. (1990) The location of binding sites on C1q for DNA. J. Immunol. 144, 3484–3488. 30. Jiang, H., Cooper, B., Robey, F.A. & Gewurz, H. (1992) DNA binds and activates complement via residues 14–26 of the C1q A chain. J. Biol. Chem. 267, 25597–25601. 31. Colliec, S., Boisson-Vidal, C. & Jozefonvicz, J. (1994) A low molecular weight fucoidan fraction from the brown seaweed Pelvetia canaliculata. Phytochemistry 35, 697–700. 32. Mulloy, B., Gee, C., Wheeler, S.F., Wait, R., Gray, E. & Bar- rowcliffe, T.W. (1997) Molecular weight measurements of low molecular weight heparins by gel permeation chromatography. Thromb. Haemost. 77, 668–674. 33. Bensimon, A., Simon, A., Chiffaudel, A., Croquette, V., Heslot, F. & Bensimon, D. (1994) Alignment and sensitive detection of DNA by a moving interface. Science. 265, 2096–2098. 34. Bensimon, D., Simon, A., Croquette, V. & Bensimon, A. (1995) Stretching DNA with a receding meniscus: experiments and models. Phys. Rev. Lett. 74, 4754–4757. 35. Illy, C., Thielens, N.M. & Arlaud, G.J. (1993) Chemical char- acterization and location of ionic interactions involved in the assembly of the C1 complex of human complement. J. Prot. Chem. 12, 771–781. 36. Arlaud, G.J., Gaboriaud, C., Thielens, N.M., Budayova-Spano, M., Rossi, V. & Fontecilla-Camps, J.C. (2002) Structural biology of the C1 complex of complement unveils the mechanisms of its activation and proteolytic activity. Mol. Immunol. 39, 383–394. 37. Kirschfink,M.,Blase,L.,Engelmann,S.&Schwartz-Albeiz,R. (1997) Secreted chondroitin sulfate proteoglycan of the human B cell lines binds to the complement protein C1q and inhibits com- plex formation of C1. J. Immunol. 158, 1324–1331. 38. Jiang, H., Burdick, D., Glabe, C.G., Cotman, C.W. & Tenner, A.J. (1994) b-Amyloid activates complement by binding to a specific region of the collagen-like domain of the C1q A chain. J. Immunol. 152, 5050–5059. 39. Bradley, A., Brooks, D., Norris-Jones, R. & Devine, D. (1999) C1q binding to liposomes is surface charge dependent and is inhibited by peptides consisting of residues 14–26 of the human C1qA chain in a sequence independent manner. Biochem. Biophys. Acta 1418, 19–30. 40. Tacnet-Delorme, P., Chevallier, S. & Arlaud, G.J. (2001) b-Amyloid fibrils activate the C1 complex of complement under physiological conditions: evidence for a site for Ab on the C1q globular regions. J. Immunol. 167, 6374–6381. 41. Budayova-Spano, M., Lacroix, M., Thielens, N.M., Arlaud, G.J., Fontecilla-Camps, J.C. & Gaboriaud, C. (2002) The crystal structure of the zymogen catalytic domain of complement C1r reveals that a disruptive mechanical stress is required to trigger activation of the C1 complex. EMBO J. 21, 231–239. 4720 B. Tissot et al. (Eur. J. Biochem. 270) Ó FEBS 2003 . Interaction of the C1 complex of Complement with sulfated polysaccharide and DNA probed by single molecule fluorescence microscopy Be ´ range ` re. whether the association of the tetramer C1r 2 – C1s 2 to C1q could interfere with the binding of C1q to DNA. We studied the binding to DNA of the C1 complex labeled