Báo cáo khoa học: Microfibril-associated glycoprotein-1 binding to tropoelastin doc

Microfibril-associated glycoprotein-1 binding to tropoelastin Multiple binding sites and the role of divalent cations Adam W. Clarke and Anthony S. Weiss School of Molecular and Microbial Biosciences, University of Sydney, NSW, Australia Microfibrils and elastin are major constituents of elastic fibers, the assembly of which is dictated by multimolecular interactions. Microfibril-associated glycoprotein-1 (MAGP- 1) is a microfibrillar component that interacts with the sol- uble e lastin precursor, tropoelastin. We describe here the adaptation of a solid-phase binding assay that defines the effect of divalent cations on the i nteractions between MAGP-1 and tropoelastin. U sing this assay, a strong cal- cium-dependent interaction was demonstrated, with a dis- sociation constant of 2.8 ± 0.3 n M , which fits a single-site binding model. Manganese a nd magnesium bestowed a weaker association, and copper did not facilitate the protein interactions. Three constructs spanning tropoelastin were used to quantify their relative contributions to calcium- dependent MAGP-1 binding. B inding to a construct span- ning a region from t he N-terminus to domain 18 followed a single-site binding model with a dissociation c onstant of 12.0 ± 2.2 n M , which contrasted with the complex bind ing behavior observed for fragments s panning domains 17–27 and domain 27 t o the C-terminus. To further elucidate binding sites around the kallikrein cleavage site of domains 25/26, MAGP-1 was p resented with constructs containing C-terminal deletions within the region. Construct M1659, which spans a r egion f rom t he N-terminus of tropoelastin to domain 26, inclusive, bound MAGP-1 with a dissociation constant of 9.7 ± 2.0 n M , which d ecreased to 4.9 ± 1.0 n M following the removal of domain 26 (M155n), thus dis- playing only half the t otal capacity to bind MAGP-1. These results demonstrate that MAGP-1 is capable of cumulative binding to distinct regions on tropoelastin, with different apparent dissociation constants and different amounts of bound protein. Keywords: MAGP-1; tropoelastin; dissociation constants; calcium; elastin. Elastic c onnective tissue has high resilience and plays an important role in tissues such as lung, skin, ligament, blood vessels, and cartilage, in particular because of the presence of elastic fibers interwoven within a complex extracellular m atrix. Elastic fibers comprise an amorphous core of cross-linked elastin and microfibrils that contain diverse macromolecules, including microfibril-associated glycoprotein-1 ( MAGP-1) whose b iological role h as yet to be elucidated [1,2]. As a microfibrillar component, MAGP-1 is spatially located to specifically interact with fibrillins and elastin [3]. MAGP-1 was first isolated from nuchal ligaments under denaturing and reduced conditions [4], and localized to beaded microfibrils [5]. Since its discovery, there has been great e ffort to determine the role played by MAGP-1 in elastic fiber assembly. M AGP-1 directly binds tropoelastin, the soluble precursor of elastin [6]. Human tropoelastin (Fig. 1 ) c ontains two major types of domains, namely (a) hydrophobic domains, rich in nonpolar amino acids, usually found in repeat sequences, and (b) hydrophilic domains, rich in Ala and Lys, which are oxidized by lysyl oxidase as a prelude to forming cross-linked i nsoluble e lastin. T he C-terminus of tropo- elastin contributes to the binding of MAGP-1 to tropoelastin [3,6]. The current studies provide an explanation for the marked decrease in interaction when tropoelastin is cleaved by plasma or pancreatic kallikrein into two fragments [7] and simplify its reportedly complex interactions [8]. Furthermore, the current work explains the incomplete blockage of MAGP-1 binding by an antibody targeted to the C-terminus of tropoelastin [3]. When calcium is removed from isolated microfibrils, gross morphological disruption occurs within the micro- fibrils [9]. Calcium is v ital to the interactions of other extracellular matrix proteins, such as the fibrillins [8–10] and the fibulins [11,12]. However, prior to this study, no work had been carried out into the effect of calcium on the interaction of tropoelastin with other e xtracellular matrix proteins. This study aims to identify, for the first time, the need for specific divalent cations in the interaction between MAGP-1 and tropoelastin. We define dissociation constants for MAGP-1 to tropoelastin in the presence of varying concentrations of divalent metal ions, demonstrate a preference for calcium, and discover that MAGP-1 binds to multiple regions on tropoelastin. Correspondence to A. Weiss, Scho ol of Molecular and Microbial Biosciences G08, University of Sydney, Sydney, NSW 2006, Australia. Fax: + 61 29351 3467, Tel.: + 61 29351 3434, E-mail: a.weiss@mmb.usyd.edu.au Abbreviations: MAGP-1, microfibril-associated glycoprotein-1. (Received 5 April 2004, revised 19 M ay 2004, accepted 7 June 2004) Eur. J. Biochem. 271, 3085–3090 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04246.x Materials and methods Human tropoelastin constructs Full-length human tropoelastin (SHELD26A), which lacks the region corresponding to exon 26A [13], SHELN-18, SHEL17-27, and SHEL27-C were prepared as previously described [8,14]. SHELN-18 spans the region of the SHELD26A gene corresponding to exons 2–18. SHEL17-27 spans the region corresponding to exons 17–27, while SHEL27-C spans the region corresponding to exon 27 to the C-terminus. The tropoelastin construct, M155n, was prepared as previously described [15] (Fig. 1). M1659 was produced as a r esult of an i nsertion of a T in to the e lastin gene at position 1659. It gives rise to a frameshift mutation, leading to a truncated form of tropoelastin containing a 21 amino acid non-native peptide sequence at t he C-terminus. All proteins w ere e xpressed r ecombinantly in Esche richia coli BL21(DE3). The purity o f each c onstruct was examined using SDS/PAGE (Fig. 2) and the concentration determined using the bicinchoninic acid protein ass ay (Pierce). Production and concentration determination of MAGP-1 MAGP-1 was purified to an extent that was previously described [8], with BSA added to aid in the stability and serve as a carrier du ring purification of MAGP-1. The identity of the resultant MAGP-1 was confirmed u sing Western blot methods, as described below. It was further tested in ligand o verlay assays, as previously described, and showed the same binding behaviour therein [ 8] (Fig. 3). The concentration of MAGP-1 was determined using previously described methods for small quantities of protein [16,17]. Western blots were performed using a T7 tag positive control (Novagen) at a standard concentration of 2.5 lgÆmL )1 . Six different amounts of the T7 tag positive control, along with MAGP-1, were loaded onto an SDS polyacrylamide gel. After separation, a Western blot was performed as described below. The resultant bands were measured for intensity using IMAGEQUANT 5.1 Software. A standard curve was constructed using the T7 tag positive control and the concentration o f M AGP-1 w as measured. The process was repeat ed in duplicate a nd an average concentration of MAGP-1 was determined. Western blotting Protein samples were first separated by SDS/PAGE [18]. Transfer to poly(vinylidene difluoride) was performed in Fig. 2. Tropelastin constructs. SDS/PAGE sh owing t he protein profile of tropoelastin constructs. The positions of the size marker proteins are indicatedinkDa. Fig. 3. Ligand overlay blot showing binding of MAGP-1 to tropoelastin. Lane 1, tropoelastin run on SDS/PAGE. Fo llowing S DS/PAG E, tropoelastin w as b lotted onto poly(vinylidene d ifluoride) for a ligand overlay assay with M AGP-1 as the s oluble ligand. Lane 2, MAGP-1- bound full-length tropoelastin. Size m arkers are i n kDa. Fig. 1. Domain structure of tropoelastin constructs. The domain structure of h uman t ropoelastin (SHELD26A) is compared with that of constructs SHELN-18, SHEL17-27, SHEL27-C, M155n and M1659. The dotted line r efers to the location of do main 26A which has been om itted from full-length tropoelastin and S HEL 17-27. Tropo - elastin i s produced naturally as a p reprot ein with a n N-terminal 26 amino acid s ignal peptide, w hich is then proteolytically removed. The construct M1659 contains a non-native 21 amino acid C-terminal extension, which arises through a frameshift mutation a t position 1659 in the elastin gene. The hydrophobic domains and lysyl oxidase crosslinking reg ions are also highlighted. 3086 A. W. Clarke and A. S. Weiss (Eur. J. Biochem. 271) Ó FEBS 2004 transfer buffer [12 m M Tris, 96 m M glycine, pH 8 .3, 20% (v/v) methanol] using a H oefer Transblot apparatus (Hoefer, Inc., San Francisco, CA, USA). Transfers w ere performed at 70 mA for 16 h at 4 °C. After the transfer, membranes were blocked in 3% (w/v) n onfat milk powder i n TBST (50 m M Tris, pH 7 .4, 150 m M NaCl, 0.1% Tween-20) for 30 min, and washed t hree times, for 10 min each wash, in TBST. The membrane was then incubated with anti-T7 horseradish peroxidase conjugate (EMD Biosciences, Inc., San Diego, CA, USA) for 30 min. After washing th ree times, for 10 min each wash, with TBST, the membrane was developed by incubation in diaminobenzidine (100 m M Tris/HCl, pH 7.5, with 4 m M diaminobenzidine, 1.7 m M NiCl 2 , and 0.012% H 2 O 2 ).Thereactionwasallowedtoproceedfor5–10min, until bands were clearly visible, then terminated by rinsing the m embrane i n distilled water. Solid-phase binding assays Solid-phase binding assays with MAGP-1 were conducted with recombinant t ropoelastin, MAGP-1, SHELN -18, SHEL17-27, SHEL27-C, M1659 and M155n. The method was a dapted from previous studies involving MAGP-1 a nd tropoelastin [8]. For all assays involving tropoelastin or tropoelastin derivatives, the tropoelastin, in NaCl/P i con- taining 2 m M EDTA, was added to wells of a microtitre plate a t a concentration o f 1 0 lgÆmL )1 andallowedtocoat overnight at 4 °C. EDTA was included t o remove all traces of divalent cations in the system. The wells were washed twice w ith N aCl/P i + EDTA, then blocked with 5 % (w/v) skim milk powder in NaCl/P i + E DTA at room tempera- ture for 2 h. After thoroughly rinsing the w ells using NaCl/ P i + EDTA, MAGP-1 was added to NaCl/P i containing 0.1% (w/v) s kim milk and various concentrations of CaCl 2 , MnSO 4 ,MgCl 2 or CuSO 4 (Table 1). To e xamine MAGP-1 binding to tropoelastin in the absence of CaCl 2 ,MAGP-1 was ad ded to the wells in the presence of 2 m M EDTA in NaCl/P i containing 0.1% (w/v) skim milk. MAGP-1 was incubated with t ropoelastin for 3 h at 37 °C. Following this step, the wells were thoroughly washed using NaCl/P i containing 0.1% (w/v) s kim milk/metal ions buffer. T7 antibody–horseradish peroxidase conjugate w as added a t a ratio of 1 : 3000 in NaCl/P i containing 0.1% (w/v) skim milk and metal ions, incubated at room temperature for 1.5 h then washed with NaCl/P i containing 0.1% (w/v) skim milk and metal ions. C olor was developed using 1 mgÆmL )1 5-aminosalicylic acid (Sigma-Aldric h) in 20 m M phosphate buffer pH 7.0, containing 0.05% (v/v) H 2 O 2 , and measured at 450 nm. This wavelength was chosen owing to the limited spectral range of the machine (Bio-Rad M odel 450 Micro- plate Reader; Bio-Rad). Negative controls for all assays were performed using 10 lgÆmL )1 BSA instead of tropo- elastin o r tropoelastin constructs. The dissociation constant (k D ) was determined using nonlinear regressional analysis ( GRAPH PAD PRISM P RO v. 4.01), assuming s ingle-site binding behavior. The linearity of response was confirmed by separate assays (Biacore International AB, Sweden). Bold curves indicate those data f or which nonlinear regression analysis was performed. All assays were repeated in quadruplicate. Results MAGP-1 binding to tropoelastin is dependent on divalent cations Solid-phase binding assays were used to determine the effect of metal ions on the binding of MAGP-1 to tropoelastin. Strong binding occurred in the presence of 1 m M calcium, with a k D of 2.8 ± 0.3 n M . This was followed by weaker binding in the presence of 1 m M manganese (k D ¼ 6.0 ± 0.6 n M )and1m M magnesium (k D ¼ 7.4±1.0n M ) (Table 1, Fig. 4). All three i ons showed binding in a Table 1. Summary o f dissociation constants for MAGP-1 binding to tropoelastin constructs. All values were determined from g raphical data with an R 2 value ¼ 0.97. V alues represent the m ean ± SD of quad- ruplicate experiments. No binding, no interaction was detected between the construct and MAGP-1 at the given concentration. Binding occurred, experiments at which an interaction was observed but whe re no dissociation co nstant c ould be determined (see the text). Construct Ion present Concentration (m M ) k D (n M ) Tropoelastin Calcium (II) 1 2.8 ± 0.3 Tropoelastin Calcium (II) 0.5 Binding occurred Tropoelastin Calcium (II) 0.1 No binding Tropoelastin EDTA 2 No binding Tropoelastin Manganese (II) 1 6.0 ± 0.6 Tropoelastin Magnesium (II) 1 7.4 ± 1.0 Tropoelastin Copper (II) 1 No binding SHELN-18 Calcium (II) 1 12.0 ± 2.2 SHEL17-27 Calcium (II) 1 Binding occurred SHEL27-C Calcium (II) 1 Binding occurred M1659 Calcium (II) 1 4.9 ± 1.0 M155n Calcium (II) 1 9.7 ± 2.0 Fig. 4. Binding of so luble MAGP-1 to tropoelastin in the pr e sence of divalent cations. Tropoelastin was coated onto t he plastic surface of multiwell plates at 1 0 lgÆmL )1 in the presence of 2 m M EDTA, and the negative control was incubated with 10 lgÆmL )1 BSA in t he presence of 2 m M EDTA. Incubation with MAGP-1 (0–35 n M )wasperformed at 37 °C i n the presence of 1 m M CaCl 2 (j), 1 m M MnSO 4 (r), 1 m M MgCl 2 (m)and1 m M CuSO 4 (.). Bound MAGP-1 was detected b y an anti-T7 immunoglobulin. Binding of MAGP-1 to BSA was negligibl e in the presence of all cations and has b een omitted from the results. Results are sh own as the me an ± SD o f quadruplicate values. Ó FEBS 2004 Multiple binding by MAGP-1 to tropoelastin (Eur. J. Biochem. 271) 3087 dose-dependent and saturating m anner. In contrast, 1 m M copper inhibited binding between t he two p roteins. Further- more, binding saturation decreased in the presence of manganese and magnesium compared with c alcium. MAGP-1 binding to tropoelastin depends on the calcium concentration Tropoelastin showed strong binding in the presence of 1m M calcium. In contrast, at concentrations of £ 0.5 m M calcium, both t he binding strength and saturation w ere substantially reduced (Fig. 5). Calcium binding was oblit- erated at or below concentrations of 0.1 m M .WhenEDTA was added, no binding was observed (Fig. 5). MAGP-1 binds to multiple tropoelastin constructs Three constructs were initially used with the aim of identifying binding sites on t ropoelastin. SHELN-18 spans a region from the N-terminus to domain 18, SHEL17-27 spans a region from domain 17 to domain 27, and SHEL27- C spans a region from domain 27 to the C-terminus (Fig. 1 ). All t hree co nstructs showed binding to MAGP-1 at different levels (Fig. 6); however, only S HELN-18 showed a b inding profile that could be m odeled to single-site binding behavior (Fig. 6 ). Using the binding curve, the k D for t he interact ion of SHELN-18 and MAGP-1 was determined to be 12.0 ± 2.2 n M (Table 1). The binding s aturation o f SHELN-18 w as significantly lower t han that of full-length tropoelastin (Fig. 6). Binding of MAGP-1 occured to constructs SHEL17-27 and SHEL27-C, which can be observed b y c omparison t o t he b ack ground binding of BSA. The constructs showed a binding saturation which was markedly lower than that of tropoelastin (Fig. 6 ), where no reliable m odel of binding behavior could be modeled to t hese fragments, p resumably as a re sult of more complex binding at weaker k D values. MAGP-1 binding is influenced by part of domain 26 MAGP-1 binding was examined using constructs M1659 (which comprises a region from the N-terminus to part of domain 26) and M155n (which spans a region from the N-terminus to domain 25) (Fig. 1). These displayed differ- ent binding profiles to MAGP-1 (Fig. 7 ). The strongest interactions o ccurred with M 1659 (k D ¼ 4.9 ± 1.0 n M ), followed by binding to M155n (k D ¼ 9.7 ± 2.0 n M ), with a commensurate loss of approximately 50% of the bound MAGP-1. This indicates the importance of either part of domain 26 or an i ntact 25–26 junction. The o verall bi nding intensities decreased with the increasing loss of the C-terminus (tropoelastin > M1659 > M155n). Discussion The binding profiles of tropoelastin to extracellular m atrix proteins are l argely unknown [1]. Although some i nter- actions between tropoelastin and MAGP-1 have been documented [3], systematic analysis o f the number of binding sites, their relative binding efficiencies and the effects of divalent metal ions have not b een examined. The current studies demonstrate a dependence o n divalent metal ions and the involvement of multiple r egions of tropoelastin. These data argue for a virtual coating of tropoelastin with MAGP-1 across much of the molecule during elastic fiber assembly [19]. Fig. 5. Binding of so luble MAGP-1 to tropoelastin in the pr esen ce of different concentrations of calcium. Tropoelastin was c oated on to the plastic surface of multiwell plates at 1 0 lgÆmL )1 in the presence o f 2m M EDTA, and the n egative control was incubated with 10 lgÆmL )1 BSA in the presence of 2 m M EDTA. Incubation with MAGP-1 (0–35 n M )wasperformedat37°C in the presence of 0.5 m M CaCl 2 (m), 0.1 m M CaCl 2 (.) and no CaCl 2 (2 m M EDTA) (j). Bound MAGP-1 was detected by an antibody against T7. Binding of M AGP- 1 t o BSA was negligible in the p resence of all calcium concen tratio ns and has been omitted from t he results. Results are expressed as the mean ± SD o f q uadruplicate values. The B SA contr ol was ne gative and has be en omitted for c larity. Fig. 6. Binding o f soluble M AGP-1 to t he tropoelastin constructs SHELN-18, SHEL17-27, and SHEL27-C. Tropoelastin constructs were coated onto the plastic surface of multiwell plates at 10 lgÆmL )1 and the negative control was incubated with 10 lgÆmL )1 BSA. Incu- bation with MAGP-1 (0–35 n M )wasperformedat37°Cinthepres- ence of 1 m M CaCl 2 . Bound MAGP-1 was detected by an anti-T7 immunoglobulin . Binding of MAGP -1 to SHELN-18 (r), SHEL17-27 (d), SHEL27-C ( m), an d non specific b inding to BSA (.)wasmon- itored by the c hange in a bsorbance at 450 nm . Results are s hown a s the m ean ± SD of quadruplicate v alues. Curve to which nonlinear regression analys is was p erformed is indicate d in b old. 3088 A. W. Clarke and A. S. Weiss (Eur. J. Biochem. 271) Ó FEBS 2004 The recombinant form of M AGP-1, used in the current study, is functionally useful because it (a) displays stronger binding to tropoelastin than other forms of MAGP-1, ( b) displays saturable b inding to defined regions on tropoelas- tin, and (c) demonstrates an absolute requirement for divalent cations. Bovine MAGP-1 is normally extracted under extreme conditions to remove it from tenacious tissues, after which i t binds tropoelastin with a k D of 260 n M [20]. The value is probably h igh a s a result of the harsh conditions (including reductive saline t reatment) under which native bovine MAGP-1 is extracted, leading t o a less function al protein. Recombinant MAGP-1, from a novel mammalian e pisomal expression system, gave a k D valueofof22±7n M [19], which a pproaches our value o f 2.8 ± 0.3 n M .Therecom- binant form of MAGP-1 is remarkably stable and displays saturable, reversible binding t o tropoelastin. The s trong k D may arise owing to a number of factors, such as (a) the increased solubility of t his form of MAGP-1 and (b) careful control over the concentrations of divalent cations in the system. No other comparable study has examined the need for divalent cations. In this study, four divalent cations were tested. Calcium enabled the tightest binding between MAGP-1 and tropoelastin. Oth er divalent cations facilita- ted d ifferent binding profiles between the two proteins, wherecationremovalledtoacessationofbinding. Accordingly, calcium p lays a dominant role in defin ing interactions at the binding sites of these two proteins. Previous studies indicated that binding occurred between the two proteins in a c ation-independent manner [3,6]. Without taking any measures t o reduce free cations in solution, it is probable t hat the use of s kim milk as a blocking reagent in previous studies may account for the presence of contaminating c ations that could contribute to binding. Although cations must have been present in previous studies, they were present in insufficient a mounts to produce a strong interaction, thus explaining the differences in k D values obtained in this work compared with those of p revious studies. H ere, the u se of a m etal ion chelator in the blocking reagents eliminated the possibility that contaminating cations were present i n the system, and ensured that the interactions observed were a direct response to those cations present in the buffer. A direct calcium-dependence was observed between tropoelastin and MAGP-1, and compares with a requirement for calcium in the binding of MAGP-1 to fibrillin-1 [8] a nd of fibrillin-1 to tropoelastin [19]. Most studies of binding with MAGP-1 have concentrated on the C-terminus of tropoelastin as the major bindin g site. An antibody targeted to the C-terminus only partly inhibited binding by MAGP-1 [3]. The current studies explain these data by showing that binding occurs to at least a single site between domains 2 and 18. In the regions spanning domain 17 to the C-terminus, m ultiple binding occurs in a complex manner that is not easily defined. It was disc overed t hat t ropoelastin digested w ith kallik- rein was unable to bind to MAGP-1 [8]. The kallikrein digest site occurs at the junction of domains 25 and 26. Theseresultssuggestthatthesiteiscrucialinthebindingof MAGP-1 to tropoelastin. We i nvestigated this region using mutated derivatives of tropoelastin. M1659 contains the major p art of domain 26, whereas M155n ends at domain 25 prior to domain 26. The results confirm that this part of domain 26 is important. M1659 showed a strong bind- ing affinity, whereas M155n exhibited a weaker binding efficiency. Although M1659 contains a non-native C-terminal region, the effect of which is unknown, these two fragments otherwise differ b y the p resence/absence of domain 26. As shown by these s tudies, t he region probably contains a binding site for MAGP-1. One of these bin ding sites is located at the junction of domains 25 and 26, while another is in the first 36 residues of domain 26. Our data present a complex binding stoichiometry f or M AGP-1 that involves multiple binding sites and differing affinities dominated by these two regions. Binding was confirmed by Biacore to be reversible (data not shown). The presence of multiple binding sites reveals that tropoelastin can present alternative binding regions to MAGP-1 during elastogenesis and presumably in the elastic fiber, and b inds multiple copies of MAGP-1 based on th e stoichiometry o f available sites. Acknowledgements The re search w as supported by g rants to A.S.W. f rom the Australian Research Council. A .W.C. is a recipient of an A ustralian P ostgraduate Research Award. References 1. Vrhovski, B. & Weiss, A.S. (1998) Biochemistry of tropoelastin. Eur. J. Biochem. 258, 1 –18. 2. Prosser, I.W. & Mecham, R.P. (1988) Re gulation of extracellular matrix accumulation: a short review of elastin biosynthesis. In Self-Assembling Architecture (Varner,J.E.,ed.),pp.1–23.Liss, New York. Fig. 7. Binding of soluble microfibril-associated glyc oprotein-1 (MAGP-1) to tropoelastin c onstructs M1659 and M155 n. Constructs were coated to the plastic surface of multiwell plates at 1 0 lgÆmL )1 and the negative c ontrol was incubated with 10 lgÆmL )1 BSA. Incubation with MAGP-1 (0–37 n M )wasperformedat37°C in the prese nce o f 1m M CaCl 2 . Bound MAG P-1 was d ete cted by a n a nti-T 7 immuno - globulin. Binding of MAGP-1 to M1659 (m), M155n (.), and non - specific binding to BSA (d) was m onitored b y the c hange in absorbance at 450 nm. Results are shown as the mean ± SD of quadruplicate values. Ó FEBS 2004 Multiple binding by MAGP-1 to tropoelastin (Eur. J. Biochem. 271) 3089 3. Brown-Augsburger, P., Broekelmann, T ., Mecham, L., Mercer, R., G ibso n, M.A., Cleary, E.G., A brams, W.R., Rosenbloom , J. & Mecham, R.P. (1994) Microfibril-associated glycopro tein bin ds to the carboxyl-terminal domain of tropoelastin and is a substrate for transglutaminase. J. Bi ol. Chem. 269, 28443–28449. 4. Gibson, M.A., Kumaratilake, J.S. & Cleary, E.G. (1989) The protein components of the 12-nanometer microfibrils of elastic and non-elastic tissues. J. Biol. Chem. 264, 4590–4598. 5. Henderson, M., Polewski, R., Fanning, J.C. & Gibson, M.A. (1996) Microfibril-asssociated g lycoprotein-1 (MAGP-1) is s peci- fically located on the beads of the beaded-filament structure for fibrillin-containing microfibrils as visualized by the rotary sha- dowing technique. J. Histochem. Cytochem. 44 , 1389–1397. 6. Mecham, R.P., Br oekelmann, T., D avis, E.C., Gibson, M.A. & Brown-Augsburger, P. (1995 ) Elastic fibr e as sembly: m acro- molecular interactions. Ciba Foundat. Sympos. 192 , 172–181, dis- cussion 1 81–184. 7. Jensen, S.A., Vrhovski, B . & Weiss, A.S. (2000 ) Domain 26 of tropoelastin plays a dominant ro le in association by c oacervation. J. Bi ol. Chem. 275, 28449–28454. 8. Jensen, S.A., Deiter, P.R., Gibson, M.A. & Weiss, A.S. ( 2001) Protein interaction studies of MAGP-1 with tropoelastin and fibrillin-1. J. Bio l. Chem. 276, 3 9661–39666. 9. Kielty, C .M. & Shuttleworth, C.A. (1993) The role o f calcium in the organization of fibrillin microfibrils. FEBS Lett. 336, 323–326. 10. Reinhardt, D.P., Mechling, D.E., Boswell, B.A., Keene, D.R., Sakai, L.Y. & Bachinger, H.P. (1997) Calcium determines the shape of fibrillin. J. Biol. C hem. 272, 7368–7373. 11. Yanagisawa, H., Davis, E. C., S tarcher, B.C., Ouchi, T., Yanagi- sawa, M ., Richardson, J .A. & Olson, E.N. ( 2002) Fibulin-5 is an elastin-binding protein essential for elastic fi bre development in vivo. Nature 415, 1 68–171. 12. Olin, A.I., Morgelin, M ., Sasaki, T., Timpl, R. , Heinegard, D. & Aspberg, A. (2001) The proteoglycans aggrecan a nd Versican form ne tworks with fibulin-2 th rough their lectin domain binding. J. Bi ol. Chem. 276, 1253–1261. 13. Martin, S.L., Vrhovski, B . & Weiss, A.S. (1995) Total s ynthesis and expression in Escherichia c oli of a gene e ncoding h uman tro- poelastin. Gene 154 , 159–166. 14. Rodgers, U.R. & Weiss, A.S. (200 4) Integrin a v b 3 binds a u nique non-RGD site near the C-terminus of h uman trop oelastin. Biochimie 86, 173–178. 15. Wu, W.J. & Weiss, A.S. (1999) Deficient coacervation of two forms of human tropoelastin associ ated with supravalvular aortic stenosis. Eur. J. Biochem. 266, 308–314. 16.Holmes,J.L.&Pollenz,R.S.(1997) Determination o f aryl hydrocarbon rec eptor nuclear transloca tor protein concentration and subcellular localization in hepatic and nonhepatic cell culture lines: develo pment of quantitative Western blotting protocols for calculation of a ryl h ydrocarbon receptor and aryl hydrocarbon receptor nuclear translocator protein in total cell lysates. Mol. Pharmacol. 52, 202–211. 17. Taylor, F .R., Ferrant, J.L., F oley, S.F., Zeng, C., Sernatinger, J., Juffras, R. & Pepinsky, R.B. (2001) Biochemical analysis of retroviral structural proteins to ide ntify and quantify r etrovirus expressed by an N S0 murine myeloma cell line. J. Biotechn ol. 84, 33–43. 18. Laemmli, U.K. (1970) C leavage of structural proteins du ring the assembly of th e head of b acteriophage T4. Na ture 227, 6 80–685. 19. Rock, M.J., Cain, S.A., Freeman,L.J.,Morgan,A.,Mellody,K., Marson, A., Shuttleworth, C.A., Weiss, A.S. & Kielty, C.M. (2004) Molecular basis of elastic fiber formation: critical inter- actions and a tropoelastin–fibrillin-1 crosslink. J. Biol. Chem. 279, 23748–23758. 20. Finnis,M.L.&Gibson,M.A.(1997) Microfibril-associated gly- coprotein-1 (MAGP-1) binds to the pepsin-resistant domain of t he alpha3 (VI) chain o f type VI collagen. J. Biol. C hem. 272, 22817– 22823. 3090 A. W. Clarke and A. S. Weiss (Eur. J. Biochem. 271) Ó FEBS 2004 . in quadruplicate. Results MAGP-1 binding to tropoelastin is dependent on divalent cations Solid-phase binding assays were used to determine the effect of metal ions on the binding of MAGP-1 to tropoelastin. Strong binding. present Concentration (m M ) k D (n M ) Tropoelastin Calcium (II) 1 2.8 ± 0.3 Tropoelastin Calcium (II) 0.5 Binding occurred Tropoelastin Calcium (II) 0.1 No binding Tropoelastin EDTA 2 No binding Tropoelastin Manganese. Microfibril-associated glycoprotein-1 binding to tropoelastin Multiple binding sites and the role of divalent cations Adam W. Clarke