Advertisement

Influences of the N700S Thrombospondin-1 Polymorphism on Protein Structure and Stability*

  • C. Britt Carlson
    Footnotes
    Affiliations
    Departments of Medicine and Biomolecular Chemistry, University of Wisconsin, Madison, Wisconsin 53706
    Search for articles by this author
  • Yuanyuan Liu
    Affiliations
    Departments of Medicine and Biomolecular Chemistry, University of Wisconsin, Madison, Wisconsin 53706
    Search for articles by this author
  • James L. Keck
    Affiliations
    Departments of Medicine and Biomolecular Chemistry, University of Wisconsin, Madison, Wisconsin 53706
    Search for articles by this author
  • Deane F. Mosher
    Correspondence
    To whom correspondence should be addressed: 4285 Medical Sciences Center, 1300 University Ave., University of Wisconsin, Madison, WI 53706. Tel.: 608-262-1576; Fax: 608-263-4969
    Affiliations
    Departments of Medicine and Biomolecular Chemistry, University of Wisconsin, Madison, Wisconsin 53706
    Search for articles by this author
  • Author Footnotes
    * This work was supported, in whole or in part, by National Institutes of Health Grant HL54462 (to D. F. M.). This work was also supported by a Shaw Foundation for Medical Research grant (to J. L. K.). Differential scanning calorimetry data were obtained at the University of Wisconsin-Madison Biophysics Instrumentation Facility, which was established with support from the University of Wisconsin-Madison, National Science Foundation Grant BIR-9512577, and National Institutes of Health Grant S10 RR13790. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
    The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1 and S2.The atomic coordinates and structure factors (code 2RPH) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).
    1 Supported by National Institutes of Health Training Grant HL07899.
Open AccessPublished:May 22, 2008DOI:https://doi.org/10.1074/jbc.M800223200
      Thrombospondins (THBSs) are multimodular, secreted proteins characterized by a signature domain comprising a unique set of 13 calcium-binding repeats flanked by epidermal growth factor (EGF)-like and lectin-like modules. A polymorphism that changes a conserved Asn to Ser at residue 700 in the most N-terminal calcium-binding repeat of THBS-1 (repeat 1C) is found in 8–10% of European populations and has been linked to increased risk of premature coronary artery disease. The Ser substitution leads to altered stability in the EGF-like and wire modules of the THBS-1 signature domain as assessed by differential scanning calorimetry carried out in 2 mm or 200 μm calcium. Studies of the melting profiles of the THBS-2 signature domain proteins with Asn or Ser at position 702 (homologous to 700 in THBS-1) revealed that the impact of the Ser allele is similar in both THBS-1 and THBS-2. Structure determination of the Ser702 THBS-2 variant in 2 mm calcium showed that repeat 1C contains two bound calcium ions as in the crystal of the Asn702 protein, including the ion that is coordinated by Asn702, and is associated with changes in conformation of repeat 1C and the adjacent EGF-like modules. The Ser substitution leads to the decreased ability of soluble THBS-2 signature domain protein to bind 4B6.13, a conformation-sensitive monoclonal antibody that recognizes an epitope in repeat 1C. These results indicate that although THBS harboring the Ser allele binds a full complement of calcium ions, repeat 1C is altered, leading to destabilization of surrounding structures.
      Thrombospondins (THBSs)
      The abbreviations used are: THBS, thrombospondin; EGF, epidermal growth factor; r.m.s. deviation, root mean square deviation; COMP, cartilage oligomeric matrix protein; ELISA, enzyme-linked immunosorbent assay; DSC, differential scanning calorimetry; TBS, Tris-buffered saline; MOPS, 4-morpholinepropanesulfonic acid.
      3The abbreviations used are: THBS, thrombospondin; EGF, epidermal growth factor; r.m.s. deviation, root mean square deviation; COMP, cartilage oligomeric matrix protein; ELISA, enzyme-linked immunosorbent assay; DSC, differential scanning calorimetry; TBS, Tris-buffered saline; MOPS, 4-morpholinepropanesulfonic acid.
      are secreted metalloglycoproteins found in organisms as diverse as Drosophila, Ciona, and tetrapods, in which there are 5 THBSs (
      • McKenzie P.
      • Chadalavada S.
      • Bohrer J.
      • Adams J.
      ,

      Adams, J. C., Tucker, R. P., and Lawler, J. (1995) in The Thrombospondin Gene Family, pp. 11–42, R. G. Landes Company, Austin, TX

      ). THBSs are organized into two groups, based on their modular structure. Group A THBSs (THBS-1 and THBS-2) form trimers and are composed of an N-terminal module, an oligomerization domain, a von Willebrand Factor type C homology module, three THBS type 1 or properdin modules, and a signature domain comprising three epidermal growth factor-like modules (EGF1, EGF2, and EGF3), a calcium-binding wire, and a C-terminal lectin-like module (Fig. 1A). Group B THBSs (THBSs 3–5) form pentamers, lack von Willebrand factor type C and properdin modules, and have an additional EGF-like module, EGF2′ (
      • Adams J.C.
      • Lawler J.
      ). The most C-terminal EGF-like module, the wire, and lectin-like module have the highest identity, 53–82% identity, when comparing Drosophila to vertebrate THBSs (
      • Adams J.C.
      • Lawler J.
      ).
      Figure thumbnail gr1
      FIGURE 1Overview of group A thrombospondin structure. A, modular structure of a THBS-1 or THBS-2 monomer including the N-terminal module (N), the oligomerization module (o), the von Willebrand Factor type C module (VWC), the three properdin-like modules (P1, P2, and P3), the signature domain comprising three EGF-like modules (EGF1, EGF2, and EGF3), the wire, and the lectin-like module. The signature domain is colored as follows: EGF1, yellow; EGF2, orange; EGF3, wheat; wire, blue; lectin-like module, magenta; glycosylation sites, green. Truncations of the THBS-1 signature domain used for DSC studies, EGF123-1, EGF3-wire-1, and 4C-1, are shown as above. B, crystal structure of Asn702 THBS-2 signature domain protein and close-up stereo view of repeat 1C. Elements are colored as in above, with the addition of: water, cyan sphere; calcium ions, red spheres; Asn702, white Corey-Pauling-Kultin (CPK) sticks; disulfide, yellow sulfur atoms. N and C termini of repeat 1C are indicated in the close-up view. Images prepared using PyMol (
      • DeLano W.L.
      ). C, sequences of wire repeat 1C of THBS-1 and THBS-2. Residues conserved between THBS-1 and THBS-2 are shown in bold. The glycosylation site in the insert (Asn708/710) is indicated by a green arrow. Black arrows indicate Trp698/700, used as a reporter in previous biophysical studies (
      • Hannah B.L.
      • Misenheimer T.M.
      • Annis D.S.
      • Mosher D.F.
      ,
      • Hannah B.L.
      • Misenheimer T.M.
      • Pranghofer M.M.
      • Mosher D.F.
      ), and Asn700/702, the site of the THBS-1 polymorphism associated with premature coronary artery disease. Residues critical for 4B6.13 recognition, Leu703 and His722, are indicated by asterisks. As determined in the Asn702 THBS-2 signature domain structure, and suggested for THBS-1, residues involved in calcium coordination are colored as follows: bidentate side chain coordination, red; single side chain coordination, purple; main chain coordination, green; coordination via a water molecule, blue.
      Crystal structures of the signature domains of THBS-1 and THBS-2 revealed that the wire is composed of 13 calcium-binding repeats: eight C-type and five N-type (
      • Kvansakul M.
      • Adams J.C.
      • Hohenester E.
      ,
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). N-type repeats are 13 or 15 residues long, whereas C-type repeats are 23 residues long. There is an insert in repeat 1C of the wire that contains an Asn-Ala-Thr glycosylation sequence in THBS-1 and THBS-2 (Fig. 1, B and C). The THBS-2 signature domain binds 30 calcium ions, with one calcium ion present in EGF2 at the EGF1-EGF2 interface, 26 sites in the wire, and three sites in the lectin-like module. The structure of the THBS-2 signature domain revealed extensive interactions among EGF2, EGF3, the wire, and the lectin-like module (Fig. 1B). Most strikingly, the wire wraps around the lectin-like module, making contact at repeats 1C and 9C. EGF2 and EGF3 also interact with the wire-lectin complex (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ).
      THBSs have diverse roles in angiogenesis, cell motility, apoptosis, cytoskeletal organization, and extracellular matrix organization (
      • Adams J.C.
      • Lawler J.
      ). Various studies have demonstrated the ability of recombinant THBS-1 proteins made in insect cells to recapitulate functions of naturally derived proteins (
      • Narizhneva N.V.
      • Byers-Ward V.J.
      • Quinn M.J.
      • Zidar F.J.
      • Plow E.F.
      • Topol E.J.
      • Byzova T.V.
      • Zwicker J.I.
      • Peyvandi F.
      • Palla R.
      • Lombardi R.
      • Canciani M.T.
      • Cairo A.
      • Ardissino D.
      • Bernardinelli L.
      • Bauer K.A.
      • Lawler J.
      • Mannucci P.
      • Miao W.M.
      • Seng W.L.
      • Duquette M.
      • Lawler P.
      • Laus C.
      • Lawler J.

      Calzada, M. J., Kuznetsova, S. A., Sipes, J. M., Rodrigues, R. G., Cashel, J. A., Annis, D. S., Mosher, D. F., and Roberts, D. D. (2008) Matrix Biology, in press

      ). There is additional interest in these proteins because of the large number of disease-associated genetic alterations occurring within the signature domain. Heterozygosity or homozygosity for the Ser700 allele of the A8831G non-synonymous single nucleotide polymorphism in the THBS-1 gene is associated with the occurrence of premature (age < 45) coronary heart disease (
      • Zwicker J.I.
      • Peyvandi F.
      • Palla R.
      • Lombardi R.
      • Canciani M.T.
      • Cairo A.
      • Ardissino D.
      • Bernardinelli L.
      • Bauer K.A.
      • Lawler J.
      • Mannucci P.
      ,
      • Topol E.J.
      • McCarthy J.
      • Gabriel S.
      • Moliterno D.J.
      • Rogers W.J.
      • Newby L.K.
      • Freedman M.
      • Metivier J.
      • Cannata R.
      • O'Donnell C.J.
      • Kottke-Marchant K.
      • Murugesan G.
      • Plow E.F.
      • Stenina O.
      • Daley G.Q.
      ). Over 100 identified mutations in THBS-5, also known as cartilage oligomeric matrix protein (COMP) (the gene is designated COMP), lead to skeletal dysplasias when present as a single copy (
      • Posey K.L.
      • Hayes E.
      • Haynes R.
      • Hecht J.T.
      ). When considered in the context of the crystal structure of the THBS-1 and THBS-2 signature domains, these genetic changes map to the wire and lectin-like modules (
      • Kvansakul M.
      • Adams J.C.
      • Hohenester E.
      ,
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ).
      The polymorphic residue in THBS-1 (N700S) associated with premature coronary artery disease is at position 10 of wire repeat 1C (Fig. 1C) near the interaction sites among EGF3, the wire, and the lectin-like module (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). Remarkably, position 10 of all 13 wire repeats of all five vertebrate THBSs is an Asn or Asp (

      Adams, J. C., Tucker, R. P., and Lawler, J. (1995) in The Thrombospondin Gene Family, pp. 11–42, R. G. Landes Company, Austin, TX

      ,
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). Previous experiments of THBS-1 constructs revealed that the Ser-containing proteins were more sensitive to heat (
      • Hannah B.L.
      • Misenheimer T.M.
      • Annis D.S.
      • Mosher D.F.
      ) and proteolysis (
      • Narizhneva N.V.
      • Byers-Ward V.J.
      • Quinn M.J.
      • Zidar F.J.
      • Plow E.F.
      • Topol E.J.
      • Byzova T.V.
      ). A study of the effects of calcium ions on THBS-1 constructs containing EGF3 and the wire revealed that a change in calcium concentration is associated with a change in the fluorescence of Trp698. This transition occurs at higher calcium concentrations and with greater cooperativity in the Ser700 protein than the Asn700 protein (
      • Hannah B.L.
      • Misenheimer T.M.
      • Pranghofer M.M.
      • Mosher D.F.
      ). The homologous residue in THBS-2, Asn702, coordinates a calcium ion through a water molecule (Fig. 1, B and 1C), a finding consistent with the altered kinetics of a calcium-induced conformational change associated with the Ser700 allele (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ,
      • Hannah B.L.
      • Misenheimer T.M.
      • Pranghofer M.M.
      • Mosher D.F.
      ). It is not known whether the Asn to Ser change in THBS-1 causes a loss of a calcium-binding site or changes after filling of the site. It is also not known how the Ser allele impacts stability and structure of specific parts of the protein, which could lead to the changes in protein function implied in the link to disease.
      In this study, we sought to answer these questions about the structural consequences of the THBS-1 N700S single nucleotide polymorphism by investigating the changes in thermal stability of the THBS-1 signature domain and investigating the homologous N702S change in the signature domain of THBS-2, for which, unlike THBS-1, there is a crystal structure (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ) and an informative conformation-sensitive monoclonal antibody (
      • Annis D.S.
      • Gunderson K.A.
      • Mosher D.F.
      ). We found that the presence of the Ser allele alters the structure and stability of wire repeat 1C and the EFG-like modules, compared with the Asn-containing proteins.

      EXPERIMENTAL PROCEDURES

      Protein Cloning—The pAcGP67.coco baculovirus transfer vector encoding a secretion signal peptide 5′ to the cloning site followed by DNA encoding a short linker and six-histidine tag was used to enable baculovirus-driven protein expression, secretion, and subsequent purification (
      • Mosher D.F.
      • Huwiler K.G.
      • Misenheimer T.M.
      • Annis D.S.
      ). Signature domain THBS-2 comprises residues 551–1172 with ADP and ARGHHHHHH N- and C-terminal tails. The Ser702 allele was introduced by PCR mutagenesis into DNA encoding the signature domain THBS-2. Correct orientation and sequence of PCR-amplified DNA were verified by sequencing. Asn700 and Ser700 THBS-1 signature domain proteins (residues 549–1170 with ADP and NAAGHHHHHH N- and C-terminal tails), 4C-1, the lectin-like module with 10 adjacent wire repeats protein (residues 775–1170 ADP and NAGHHHHHH N- and C-terminal tails), EGF3-Wire-1, the third EGF-like module with the calcium-binding wire protein, (described previously (
      • Hannah B.L.
      • Misenheimer T.M.
      • Annis D.S.
      • Mosher D.F.
      )), and EGF123-1, the three tandem EGF-like modules protein (residues 550–690 with ADP and ALELVPRGSAAGHHHHHH N- and C-terminal tails) were constructed in a similar fashion.
      Expression and Purification of Recombinant Proteins—The Asn702 and Ser702 THBS-2 proteins were expressed by infecting High Five insect cells in SF900II serum-free medium (Invitrogen) at 27 °C with high titer virus (>1 E108 plaque-forming units/ml) at a multiplicity of infection of ∼5. Conditioned medium was collected ∼65 h post-infection. Histidine-tagged proteins were purified from the medium in the presence of 2 mm CaCl2 using Ni2+-nitrilotriacetic acid resin (Qiagen) and, if necessary, an anion-exchange HiTrap Q HP column (GE Healthcare), as described previously (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). Protein concentration was determined using absorbance at 280 nm less the absorbance at 320 nm and divided by the calculated extinction coefficient of 1.24 ml mg–1 cm–1 (THBS-2 proteins), 1.2 ml mg–1 cm–1 (THBS-1 signature domain proteins), 1.33 ml mg–1 cm–1 (4C-1), 2.42 ml mg–1 cm–1 (E123-1), or 0.81 ml mg–1 cm–1 (EGF-Wire-1) (
      • Mach H.
      • Middaugh C.R.
      • Lewis R.V.
      ).
      Differential Scanning Calorimetry (DSC)—DSC experiments were performed at the University of Wisconsin Biophysics Instrumentation Facility using a Microcal VP-differential scanning calorimeter equipped with Origin 7 software. Proteins (8.3–25 μm) were dialyzed into 10 mm MOPS, 150 mm NaCl, 2 mm or 200 μm CaCl2, pH 7.5. Scans were conducted from 15 to 95 °C at a rate of 60 °C h–1, except EGF123-1, which was scanned from 15 to 85 °C. Appropriate buffer scans were used for reference subtraction and data were normalized using protein concentrations, including an estimated 2% dilution upon introduction into the DSC sample cell. Area-fitting analyses were performed with the DSC Origin 7 software package. Manual-fit baselines were subtracted from sample scans that were corrected for reference sample and concentration. This curve was then fitted to a non-2-state model with various numbers of peaks. Fits were refined using several iterations until the χ2 value stabilized.
      Enzyme-linked Immunosorbent Assay (ELISA)—Asn702 and Ser702 THBS-2 signature domain protein were coated onto 96-well microtiter plates at 10 μgml–1 in Tris-buffered saline (TBS) containing 10 mm Tris-HCl, 150 mm NaCl, pH 7.4, plus 2 mm CaCl2. The plates were blocked with 5% bovine serum albumin in TBS plus 0.05% Tween-20 (TBST). The monoclonal antibody 4B6.13 (
      • Annis D.S.
      • Gunderson K.A.
      • Mosher D.F.
      ) was diluted in TBST containing various concentrations of CaCl2 and then incubated with the plate for 2 h. Alkaline phosphatase-conjugated secondary antibody was added for a 1-h incubation. Sigma 104 AP substrate at 1 mg ml–1 in TBS, pH 9.0, was added to each well and color development was monitored at 405 nm. As a control, we used polyclonal rabbit anti-human THBS-2 antibodies (
      • Tooney P.A.
      • Sakai T.
      • Sakai K.
      • Aeschlimann D.
      • Mosher D.F.
      ) to confirm that antigen protein was adsorbed to the plate and present at similar concentrations. The proteins were tested in competition ELISA as described (
      • Annis D.S.
      • Gunderson K.A.
      • Mosher D.F.
      ) to assess if soluble antigen was able to compete with surface-adsorbed antigen for binding of the monoclonal antibody. All ELISA experiments were repeated, each with triplicate data points, on multiple occasions. Data are expressed as the mean of the different experiments ± S.E.
      Protein Crystallization—The Ser702 THBS-2 signature domain was crystallized using the hanging drop vapor diffusion method at room temperature. Protein was first exchanged into 5 mm MOPS, pH 7.5, 2 mm CaCl2, 0.1 m sodium acetate. Protein was then concentrated to 7–10 mg ml–1 using Ultrafree-0.5 centrifugal filter units (Millipore). Drops were made from 0.9 μl of protein mixed with 0.9 μl of mother liquor (100 mm Tris-Cl, pH 8.5–9.0, 200 mm sodium acetate, 30% polyethylene glycol 4000) and 0.1 μl of 1 m glycine. Crystal formation was facilitated by streak seeding from Asn702 THBS-2 signature domain crystals (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). Prior to data collection, crystals were stabilized by transfer to a cryoprotectant solution of the above mother liquor conditions supplemented with 25% ethylene glycol and then frozen directly in liquid nitrogen.
      Protein Crystal Data Collection and Analysis—Data were collected using a Proteum CCD detector with x-rays generated by a Microstar rotating anode (Bruker AXS, Madison, WI). Data were indexed and scaled using HKL2000 (

      Otwinowski, Z., and Minor, W. (1997) in Methods in Enzymology (Carter, C. W., and Sweet, R. M., eds) Vol. 276 A, pp. 307–326, Academic Press, New York

      ). The structure was solved to 2.9-Å resolution using AMORE (
      • Navaza J.
      ) for molecular replacement with the Asn702 THBS-2 signature domain protein as a search model (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ) (Table 1). Refinement with REFMAC5 (
      CCP4
      ) produced interpretable 2FoFc and FoFc electron density maps. The model was improved by rounds of refinement with REFMAC5 and manual rebuilding using O (
      • Jones T.A.
      • Zou J.Y.
      • Cowan S.W.
      • Kjeldgaard M.
      ) and Coot (
      • Emsley P.
      • Cowtan K.
      ) to a final Rfactor of 22.7% (Rfree = 27.7%, refer to Table 1). There was one molecule per asymmetric unit. 73.7% of all residues fall in the most favored Ramachandran category, with 23.8% in the allowed category, 2.1% in the generously allowed category, and 0.4% in the disallowed category. Superpose and root mean squared analyses of Cα carbons were calculated using the CCP4 suite of programs (
      CCP4
      ).
      TABLE 1Crystallography statistics
      Data collection
      Space groupI222
      Cell dimensions
       a, b, c (Å)93.89 122.65 155.38
       α, β, γ (°)90.00 90.00 90.00
      Wavelength (Å)1.54
      Resolution (Å)80-2.9 (2.97-2.90)
      Rsym0.137 (0.372)
      I/σI9.7 (3.0)
      Completeness (%)98.1 (99.6)
      Redundancy6.8 (6.8)
      Refinement
       Resolution (Å)20-2.9
       Cut-off (σ(F))None used
       No. reflections18,868
      Rwork/Rfree22.7/27.7
      B-factors
       Protein37.9
       Ligand/ion50.0
       Water34.0
       Wilson63.2
      R.m.s. deviations
       Bond lengths (Å)0.006
       Bond angles (°)1.052
      Coordinates—Model coordinates and structure factors have been deposited in the Protein Data Bank (code 2RHP).

      RESULTS

      The Polymorphism Leads to Altered Thermal Stability of the EGF-like Modules and the Wire—To examine the impact of the Ser allele on stability of the signature domain and its specific parts, we examined the thermal melting profiles of THBS-1 and THBS-2 proteins having Asn or Ser at the position of the THBS-1 polymorphism (700 of THBS-1 and 702 of THBS-2).
      The DSC profile of the Asn700 THBS-1 signature domain in 2 mm calcium had four apparent melting events at 46.1, 53.7, 72.1, and 88.6 °C (Fig. 2A). The first two peaks were not reversible upon cooling and reheating, whereas the second two peaks were reversible (not shown). To identify the origins of these peaks, we studied various truncations of the THBS-1 signature domain (Fig. 2A). A construct comprising the three tandem EGF-like modules of THBS-1 (EGF123-1) melted reversibly with one peak at 68.0 °C. The EGF3-wire construct (EGF3-wire-1) melted reversibly with one peak at 87 °C. Because secretion of the lectin-like module requires a portion of the wire that mediates the wire repeat 9C-lectin interaction (
      • Kvansakul M.
      • Adams J.C.
      • Hohenester E.
      ), we examined a construct containing the lectin-like module and portion of the calcium-binding wire that includes repeats 4C to 13C (4C-1), including the wire repeat 9C-lectin interface (Fig. 1A). This protein melted with a non-reversible peak at 47.5 °C, followed by the beginning of a second peak at >90 °C, which was reversible. Alignment of the DSC curves (Fig. 2A) suggests that the first or second peak of the signature domain likely represents the melting of the wire repeat 9C-lectin interface, and the third and fourth peaks likely represent the melting of the EGF-like modules and the wire. Based on previous studies of the THBS-2 signature domain constructs (
      • Misenheimer T.M.
      • Hannah B.L.
      • Annis D.S.
      • Mosher D.F.
      ) (see below), we suspect that the other of the first two peaks represents melting of the wire repeat 1C-lectin interface.
      Figure thumbnail gr2
      FIGURE 2Differential scanning calorimetry. A, thermal melting curves for THBS-1 constructs in 2 mm calcium. Proteins are shown as follows: EGF123-1 (solid with triangles), Asn700 EGF3-wire-1 (solid with squares), 4C-1 (solid with circles), Asn700 signature domain (solid), and Ser700 THBS-1 signature domain (dashed). B, thermal melting curves for Asn700 THBS-1 signature domain (solid) and the Ser700 THBS-1 signature domain (dashed) in 200 μm calcium. C, thermal melting curves for the Asn702 THBS-2 signature domain (solid) and Ser702 THBS-2 signature domain (dashed). D, area-fitting curves for the Asn702 THBS-2 signature domain fit as four events (χ2 = 7.4 × 104,Tm1–4 = 51.0, 53.1, 78.8, and 82.7 °C). E, area-fitting curves for Ser702 THBS-2 signature domain fit as four events (χ2 = 7.5 × 104,Tm1–4 = 50.0, 51.0, 66.5, and 78.8 °C). A–E, bar = 10 kcal/mol deg.
      Compared with the Asn700 THBS-1 protein, the Ser700 THBS-1 protein also had two non-reversible melting events at 47.3 and 53.8 °C, but instead of the 72 and 88.6 °C peaks, there was a broad two-part reversible peak from 75 to 95 °C (Fig. 2A). Thus, the reversible high temperature melting events attributed to the EGF-like and wire modules are impacted by the amino acid substitution, whereas the non-reversible low temperature events are not.
      Previous studies with a THBS-1 construct comprising EGF3 and the wire demonstrated that the Ser allele is more susceptible to heat denaturation in 200 μm calcium, compared with the Asn-containing protein (
      • Hannah B.L.
      • Misenheimer T.M.
      • Annis D.S.
      • Mosher D.F.
      ). To examine if this is also seen in the signature domain protein, the thermal stability of Asn- and Ser-containing proteins were examined at 200 μm calcium, using DSC. At 200 μm calcium, the Asn700 protein melted with two apparent peaks at 43.5 and 69.4 °C. The Ser700 protein melted with an apparent three-part peak from 40 to 70 °C (Fig. 2B). Lowering the calcium concentration to 200 μm, therefore, lowers the temperatures of the reversible transitions and, as in 2 mm calcium, the Ser700 protein was more sensitive to heating than the Asn700 protein.
      The four peaks seen in the THBS-1 proteins contrasts with DSC studies on the signature domain of THBS-2, which melts with two peaks (
      • Misenheimer T.M.
      • Hannah B.L.
      • Annis D.S.
      • Mosher D.F.
      ,
      • Misenheimer T.M.
      • Mosher D.F.
      ) (Fig. 2C). The first peak (non-reversible) of the THBS-2 protein has been assigned to the melting of the lectin-wire interfaces and the second peak (reversible) has been assigned to the melting of the wire itself (
      • Misenheimer T.M.
      • Hannah B.L.
      • Annis D.S.
      • Mosher D.F.
      ,
      • Misenheimer T.M.
      • Mosher D.F.
      ). To test the possibility that the two THBS-2 peaks are each compound transitions comprising two melting events, we used area-fitting analyses assuming a non-two-state model based on the nonreversible status of the low temperature events. We fit the Asn702 THBS-2 melting profiles to a four-event model and a two-event model. The Asn702 THBS-2 curve modeled as two events (52.2 and 81.1 °C) had a χ2 of 8.5 × 105 (not shown), whereas this curve had a χ2 value of 7.4 × 104 for a four-event model (51.0, 53.1, 78.8, and 82.7 °C) (Fig. 2D), compatible with the hypothesis that the transitions seen for both THBS-1 and THBS-2 proteins are reporting four (or more) events.
      The DSC profile of the THBS-2 signature domain with the homologous Asn to Ser change, N702S, also fit best modeled as four events. When modeled as two events (50.9 and 77.6 °C), the Ser702 THBS-2 curve had a χ2 value of 2.3 × 106 (not shown) that decreased to 7.5 × 104 when modeled as four events (50.0, 51.0, 66.5, and 78.8 °C) (Fig. 2E). However, the melting events are altered in the Ser702 protein, compared with the Asn702 protein (Fig. 2C). The largest changes are seen in the reversible high temperature events where the Asn702 82.7 °C event is shifted to 78.8 °C in the Ser702 protein, and the Asn702 78.8 °C event is shifted to 66.5 °C in the Ser702 protein. Smaller changes are seen in the nonreversible low temperature events, where the Asn702 53.1 °C event is shifted to 51.0 °C in the Ser702 protein and the Asn702 51.0 °C event is shifted to 50.0 °C in the Ser702 protein. Thus, the major changes in the DSC pattern due to Ser702 are of the reversible high temperature events, and as with the THBS-1 protein, indicate changes in stability of the EGF-like and wire modules.
      Mutation of Asn to Ser Does Not Lead to a Loss of Calcium—A possible structural explanation for observed changes in protein stability in the Ser-containing proteins is that the Ser700 allele and Ser702 mutant are associated with a loss of a bound calcium ion. Residue 700/702 is located at position 10 of the consensus sequence found in both C-type and N-type wire repeats (Fig. 1C). Asp or Asn are invariably present at this position, and the residue coordinates calcium directly or through a water, as revealed in crystal structures of THBS-1 and THBS-2 signature domain proteins (
      • Kvansakul M.
      • Adams J.C.
      • Hohenester E.
      ,
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). Asn702 coordinates calcium through a water molecule (Fig. 1, B and C) (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). The 76% identity between the sequences of THBS-1 and THBS-2 signature domain proteins and the changes in kinetics of the conformational transition induced by calcium and reported by adjacent Trp698 in Ser700 THBS-1 proteins (
      • Hannah B.L.
      • Misenheimer T.M.
      • Pranghofer M.M.
      • Mosher D.F.
      ) argue for a similar calcium-binding role for Asn700 in THBS-1 (Fig. 1C).
      To examine the impact of the N700S polymorphism on calcium coordination, we first attempted to study the Asn700 and Ser700 THBS-1 signature domain proteins by crystallography. Our attempts to crystallize the Asn700 signature domain of THBS-1, including the C992S derivative used to stabilize and crystallize a portion of the THBS-1 signature domain (
      • Kvansakul M.
      • Adams J.C.
      • Hohenester E.
      ), were unsuccessful. We therefore addressed this question by solving the crystal structure of the Ser702 THBS-2 signature domain in 2 mm calcium. The Ser702 protein crystallized under the same conditions used for the Asn702 protein (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ) used streak seeding of the Asn702 crystals to nucleate growth. Ser702 crystals grew to a smaller size than Asn702 protein, and were solved at a maximum resolution of 2.9 Å (Table 1), as compared with 2.6 Å for the Asn702 protein (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ).
      Electron density for all 30 of the calcium ions, including the ion coordinated by Asn702 (Figs. 3, A and B) are present in the Ser702 variant, indicating that the Asn to Ser amino acid change does not result in a loss of calcium coordination. The Ser702 side chain is present in the same space occupied by the Asn702 side chain, although strong density for the terminal oxygen of Ser702 is lacking, which indicates a dynamic structure for the residue. The electron density for the coordinating water molecule seen in the Asn702 structure is not visualized in the Ser702 structure. Bond lengths and geometries between the calcium and remaining coordinating residues (Asp694, Asp696, Asp698, Trp700, and Asp718) are very similar to those of the Asn702 protein.
      Figure thumbnail gr3
      FIGURE 3Electron density and superposition analysis of Asn702 and Ser702 THBS-2 signature domain proteins. A, stereo image of 2FoFc electron density map (contoured at 1.5σ and colored blue) for Asn702 and nearby residues. B, stereo image of 2FoFc electron density map (contoured at 1.5σ and colored blue) for Ser702 and nearby residues. Calcium ions are shown as gray spheres and water molecules are shown as yellow spheres. Images were prepared using CCP4 Molecular Graphics (
      • Potterton L.
      • McNicholas S.
      • Krissinel E.
      • Gruber J.
      • Cowtan K.
      • Emsley P.
      • Murshudov G.N.
      • Cohen S.
      • Perrakis A.
      • Noble M.
      ). C and D, superpose analysis was performed on Cα atoms from the entire protein, residues 551–1171. C, overlay of wire repeat 1C and nearby residues. Calcium ions are shown as spheres. Asn/Ser702 and residues responsible for the 4B6.13 epitope, Leu703 and His722, are shown as sticks. Asn702 THBS-2 signature domain is shown in blue/light blue and Ser702 THBS-2 signature domain is shown in red/pink. Wire repeat 1C is shown in blue and red. D, r.m.s. deviation analysis of the difference between the Asn702 and Ser702 THBS-2 signature domain main chain atoms. Large red dotted box comprises wire module residues and the small red dotted box comprises wire repeat 1C residues.
      The Ser702 Mutation Leads to Altered Structure—Superposition of the Asn702 and Ser702 THBS-2 signature domain structures showed an overall root mean square deviation (r.m.s. deviation) value of 0.31 Å for all 621 Cα atoms (Fig. 3, C and D). This analysis revealed that repeat 1C (36 Cα atoms) has a modestly higher r.m.s. deviation value than the entire wire module (264 Cα atoms) (rmsdwire1C = 0.34 Å, rmsdwire = 0.29 Å). The lectin-like module (215 Cα atoms) had an overall r.m.s. deviation of 0.18 Å, in contrast to the EGF-like modules (142 Cα atoms), which had an overall r.m.s. deviation of 0.40 Å.
      Hot-spots of differences lay at the EGF1-EGF2 and EGF2-wire interfaces (supplemental Fig. S1). Specifically, at the EGF2-wire interface, the Ala603–Pro608 loop, which extends down to contact the wire repeat 10N-13C hairpin turn, shows substantial movement, as evidenced by relatively high r.m.s. deviation values (average 0.77 Å) and weak electron density in the Ser702 crystal (supplemental Fig. S2). In addition, there were other distant sites where r.m.s. deviation values were relatively high (Fig. 3D). All of these residues are surface-accessible (supplemental Fig. S1), and some contact with symmetry-related molecules in the crystal lattice. No increases in r.m.s. deviation values were observed at sites of intermodular interactions between the wire and lectin-like modules.
      Because residues in wire repeat 1C make close contacts with a neighboring protein molecule in crystals of Asn702 and Ser702 proteins, which could influence conformation in the crystal, we probed the structures of adsorbed and soluble THBS-2 signature domains with 4B6.13, a conformation-sensitive monoclonal antibody that requires calcium and the presence of EGF3, the wire, and the lectin-like module of human THBS-2 (
      • Annis D.S.
      • Gunderson K.A.
      • Mosher D.F.
      ) for binding. This antibody recognizes residues Leu703 and His722, which are proximal to the site of the N702S amino acid change (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). The side chains of Leu703 and His722 are slightly different in the two crystals (Fig. 3C). Both the Cα atom and the side chain of Leu703 are shifted in the Ser702 THBS-2 structure relative to the Asn702 THBS-2 structure. The Cα atom of His722 of Ser702 THBS-2 overlays well with the Asn702 THBS-2 His722, but the side chain is rotated slightly. The Leu703Cβ-His722Cβ distance is 13.9 Å, instead of the 13.5 Å distance seen in the Asn702 protein.
      As previously described (
      • Annis D.S.
      • Gunderson K.A.
      • Mosher D.F.
      ), 4B6.13 binding to Asn702 protein transitioned from 0 to 100% binding within a range of 140–200 μm calcium with a midpoint at 150 μm. In contrast, the Ser702 protein did not bind 4B6.13 at any calcium concentration tested (Fig. 4A). Direct ELISAs performed with polyclonal antibodies confirmed similar levels of Asn702 and Ser702 adsorption to the ELISA plates (not shown). To avoid possible effects in the direct ELISA due to altered protein conformation upon adsorption to the ELISA plates, competition ELISAs were performed in 2 mm calcium. The Asn702 signature domain competed with recombinant full-length THBS-2 for binding to 4B6.13, with half-inhibition at 0.1 μm (Fig. 4B). The Ser702 signature domain competed with recombinant full-length THBS-2 for 4B6.13 binding minimally and only at high concentrations. For 20% inhibition, greater than 100-fold higher concentrations of the Ser702 protein than the concentration of the Asn702 signature domain protein was required (Fig. 4B). The fact that the Ser702 THBS-2 protein does compete for 4B6.13 binding, however, suggests that there may be a subpopulation (∼1%) of the protein that is in the conformation recognized by 4B6.13.
      Figure thumbnail gr4
      FIGURE 4Antibody 4B6.13 binds the Ser702 THBS-2 signature domain poorly. Asn702 THBS-2 signature domain is shown as a solid line and the Ser702 THBS-2 signature domain is shown as a dashed line. A, direct ELISA, the antigen was plated in 2 mm calcium. The initial wash and blocking were in the absence of calcium and all subsequent washes and antibody incubations were at the indicated calcium concentrations. Values are expressed as the mean ± S.E. of two experiments. B, competition ELISA, competition ELISAs were preformed in the presence of 2 mm calcium. 4B6.13 and Asn702 or Ser702 THBS-2 signature domain proteins were incubated together prior to addition to plates coated with full-length recombinant THBS-2. The amount bound was quantified and compared with the absence of competitor (control). Values are expressed as mean ± S.E. of four experiments.

      DISCUSSION

      The postulated roles of THBSs are many, yet the functions of these unusual calcium-binding proteins remain enigmatic. Knock-out mice have been developed for Thbs-1 (
      • Lawler J.
      • Sunday M.
      • Thibert V.
      • Duquette M.
      • George E.L.
      • Rayburn H.
      • Hynes R.
      ), Thbs-2 (
      • Kyriakides T.R.
      • Zhu Y.-H.
      • Smith L.T.
      • Bain S.D.
      • Yang Z.
      • Lin M.T.
      • Danielson K.G.
      • Iozzo R.V.
      • LaMarca M.
      • McKinney C.E.
      • Ginns E.I.
      ), Thbs-3 (
      • Hankenson K.D.
      • Hormuzdi S.G.
      • Meganck J.A.
      • Bornstein P.
      ), and COMP (THBS-5) (
      • Svensson L.
      • Aszodi A.
      • Heinegard D.
      • Hunziker E.B.
      • Reinholt F.P.
      • Fassler R.
      • Oldberg A.
      ). In addition, Thbs-4 knock-out mice have been produced commercially. Each of the knock-out strains has only slight abnormalities, and all strains survive to adulthood and reproduce. The mild phenotypes seen with these animals lacking individual THBSs and the observations that mutations or polymorphisms of THBSs are associated with disease in heterozygotes suggest that the role of genetically altered THBS-1 and THBS-5/COMP in diseases is a result of having disrupted THBSs rather than an absence of THBS activities.
      We used the THBS signature domain to study the effects of an Asn to Ser amino acid change resulting from the premature coronary artery disease-associated polymorphism of THBS-1. The Asn700 THBS-1 signature domain melts with four peaks in DSC. Using smaller constructs of the signature domain, we attributed each of these peaks to the melting of specific parts of the signature domain. EGF123-1 reversibly melts at 68.0 °C and likely corresponds to the reversible peak seen at 72.1 °C in the complete signature domain. Past studies have shown that the addition of EGF-like modules can stabilize the wire, thus increasing the melting temperature in the longer protein constructs (
      • Misenheimer T.M.
      • Hannah B.L.
      • Annis D.S.
      • Mosher D.F.
      ,
      • Misenheimer T.M.
      • Mosher D.F.
      ). Deconstruction of the THBS-2 DSC profile indicates that the THBS-2 signature domain melts with two compound peaks, and suggests that the first THBS-2 peak corresponds to the first two peaks of THBS-1, whereas the second THBS-2 peak corresponds to the second two peaks of THBS-1. The four peaks seen in the THBS-1 signature domain and deconvoluted scans of THBS-2 were assigned as follows: peak 1 or 2, repeat 1C-lectin interface; peak 2 or 1, repeat 9C-lectin interface; peak 3, EGF123 complex; peak 4, the wire.
      The presence of Ser700 in the THBS-1 signature domain leads to altered shapes of the high-temperature transition peaks at both 2 mm and 200 μm calcium, indicating changes in the stability of EGF123 and the wire at both a calcium concentration typical of the extracellular matrix and a low calcium concentration (
      • Breitwieser G.E.
      ). Introduction of the Ser702 mutation into the THBS-2 signature domain results in a similar disruption of the melting of the EGF-like and wire modules. This analysis shows that the introduction of the Ser allele results in a local disruption of structure in the wire leading to disruption in the stability of the EGF-like modules and the wire.
      Asn702 coordinates calcium through a water molecule (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). The presence of the same calcium ion in the crystal structure of the Ser702 protein may be favored by the protein-protein interactions that drive crystallization. Is this calcium ion lost from the signature domain in a solution containing 2 mm calcium? Solution studies using a range of truncations of the signature domain of THBS-1 showed identical quenching of Trp698 upon addition of 2 mm calcium to the Asn700 and Ser700 proteins, although the Ser700 protein exhibited differences in the titration in the range of 50–150 μm calcium (
      • Hannah B.L.
      • Misenheimer T.M.
      • Annis D.S.
      • Mosher D.F.
      ,
      • Hannah B.L.
      • Misenheimer T.M.
      • Pranghofer M.M.
      • Mosher D.F.
      ). The quenching, which likely is tied to the Trp698 main chain coordination of the same calcium ion that is coordinated by the Asn700 side chain (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ), is highly cooperative with a Hill coefficient of 3.5–5.0 (
      • Hannah B.L.
      • Misenheimer T.M.
      • Annis D.S.
      • Mosher D.F.
      ). Effects of calcium on a set of truncations indicate that adjacent repeats 2N, 3C, and 4C interact allosterically with the repeat 1C (
      • Hannah B.L.
      • Misenheimer T.M.
      • Pranghofer M.M.
      • Mosher D.F.
      ). The Trp698 quenching experiments and the crystallographic results with the THBS-2 constructs, therefore, are consistent in indicating that the expected two calcium ions bind to wire repeat 1C at calcium concentrations approaching 2 mm.
      Disruption of 4B6.13 binding to the Ser702 protein and differences between the Asn702 and Ser702 THBS-2 signature domain crystal structures indicated a change in local structure upon introduction of the polymorphic allele. Residue 702 is located next to Leu703, which along with His722, is required for formation of the epitope for 4B6.13 (
      • Annis D.S.
      • Gunderson K.A.
      • Mosher D.F.
      ). Because of the proximity of residue 702 to the 4B6.13 epitope comprising residues 703 and 722, the decreased ability of the Ser702 THBS-2 protein to bind 4B6.13 could conceivably be due to a contribution of the Asn702 side chain to the 4B6.13 epitope. However, although part of the Asn main chain is surface accessible, the side chain is buried in the structure of the calcium-replete THBS-2 signature domain (
      • Carlson C.B.
      • Bernstein D.A.
      • Annis D.S.
      • Misenheimer T.M.
      • Hannah B.-l.A.
      • Mosher D.F.
      • Keck J.L.
      ). Thus, we think that Asn702 does not contribute directly to the epitope, but changes the conformation of Leu703 and His722 as shown in Fig. 2C. Supporting this hypothesis, preliminary studies of a THBS-2 signature domain variant in which Leu697 of repeat 1C, which is also not on the surface, is changed to Pro, thus mimicking the L272P mutation in THBS-5/COMP associated with pseudoachondroplasia (
      • Deere M.
      • Sanford T.
      • Francomano C.A.
      • Daniels K.
      • Hecht J.T.
      ), show that this amino acid change also leads to a dramatic decrease in 4B6.13 binding (data not shown).
      Although crystal lattice contacts may suppress or enhance effects due to the Asn to Ser amino acid change, we can make tentative conclusions based on comparing the modest resolution Asn702 and Ser702 THBS-2 signature domain structures. Relatively high r.m.s. deviation values point to alterations in individual surface residues throughout the protein, particularly in wire repeat 1C and the EGF-like modules. Of note, the N700S polymorphism is immediately adjacent to the 13-residue insert that is found in repeat 1C of THBS-1 and THBS-2 (Fig. 1, B and C). The insert interacts extensively with EGF3.
      The changes seen in surface residues and in the stability of the EGF-like modules shown by DSC suggest that long-range effects are propagated N terminally. This may extend outside the signature domain, and result in changes in the heparin-binding properties of Asn700 and Ser700 THBS-1, which are believed to be occurring at the N terminus of the molecule (
      • Narizhneva N.V.
      • Byers-Ward V.J.
      • Quinn M.J.
      • Zidar F.J.
      • Plow E.F.
      • Topol E.J.
      • Byzova T.V.
      ). A possible mechanism for increased risk for coronary artery disease, therefore, is that structural changes caused by the Ser700 allele impact the entire protein and its function. The fact that the Ser700 allele has a prevalence of 8–10% of Europeans (
      • Zwicker J.I.
      • Peyvandi F.
      • Palla R.
      • Lombardi R.
      • Canciani M.T.
      • Cairo A.
      • Ardissino D.
      • Bernardinelli L.
      • Bauer K.A.
      • Lawler J.
      • Mannucci P.
      ,
      • Boekholdt S.M.
      • Trip M.D.
      • Peters R.J.G.
      • Engelen M.
      • Boer J.M.A.
      • Feskens E.J.M.
      • Zwinderman A.H.
      • Kastelein J.J.P.
      • Reitsma P.H.
      ) suggests, however, that earlier in the history of this population, such structural and functional alterations conferred an evolutionary advantage.
      The current results provide insight into the pathology of the THBS-5/COMP mutations. Many of the THBS-5/COMP mutations change residues that directly coordinate calcium (
      • Posey K.L.
      • Hayes E.
      • Haynes R.
      • Hecht J.T.
      ). Structural studies using THBS-5/COMP constructs containing disease-associated mutations have illustrated the diversity of effects on protein structure and calcium-binding induced by the mutations (
      • Kleerekoper Q.
      • Hecht J.T.
      • Putkey J.A.
      • Chen H.
      • Deere M.
      • Hecht J.T.
      • Lawler J.
      • Hou J.
      • Putkey J.A.
      • Hecht J.T.
      • Maddox B.K.
      • Mokashi A.
      • Keene D.R.
      • Bachinger H.P.
      • Thur J.
      • Rosenberg K.
      • Nitsche D.P.
      • Pihlajamaa T.
      • Ala-Kokko L.
      • Heinegard D.
      • Paulsson M.
      • Maurer P.
      ). As previously stated, Asn700 of the THBS-1 protein and Asn702 of THBS-2 occurs at position 10 in repeat 1C of the wire. Interestingly, there are three mutations at position 10 in THBS-5/COMP (D310V in repeat 2N and D479H/D479Y in repeat 11C) associated with disease (
      • Kennedy J.
      • Jackson G.
      • Ramsden S.
      • Taylor J.
      • Newman W.
      • Wright M.J.
      • Donnai D.
      • Elles R.
      • Briggs M.D.
      ). Similar to Asn702, the THBS-2 residues at these positions, Asp738 and Asp907, coordinate a calcium ion through a water molecule. However, Asp738 and Asp907 also coordinate an additional calcium ion; Asp738 provides main chain coordination for the calcium ion found between repeats 2N and 3C, and Asp907 provides side chain coordination for the calcium ion found between repeats 10N and 11C. The THBS-5/COMP mutations result in severe phenotypes leading to multiple epiphyseal dysplasia or pseudoachrondroplasia skeletal dysplasias. The discrepancy between the presumed advantageous effect of the Ser700 THBS-1 polymorphism (in wire repeat 1C) that led to its prevalence in European populations and the disease caused by the mutation at homologous residues in wire repeats 2N and 11C in THBS-5/COMP may be because the polymorphic residue is involved in the coordination of only one calcium ion, whereas the mutated THBS-5/COMP residues coordinate two calcium ions.

      Acknowledgments

      We thank Doug Annis for help with many parts of this project, including much of the production of THBS-1.

      Supplementary Material

      References

        • McKenzie P.
        • Chadalavada S.
        • Bohrer J.
        • Adams J.
        BMC Evol. Biol. 2006; 6: 33
      1. Adams, J. C., Tucker, R. P., and Lawler, J. (1995) in The Thrombospondin Gene Family, pp. 11–42, R. G. Landes Company, Austin, TX

        • Adams J.C.
        • Lawler J.
        Int. J. Biochem. Cell Biol. 2004; 36: 961-968
        • Kvansakul M.
        • Adams J.C.
        • Hohenester E.
        EMBO J. 2004; 23: 1223-1233
        • Carlson C.B.
        • Bernstein D.A.
        • Annis D.S.
        • Misenheimer T.M.
        • Hannah B.-l.A.
        • Mosher D.F.
        • Keck J.L.
        Nat. Struct. Mol. Biol. 2005; 12: 910-914
        • Narizhneva N.V.
        • Byers-Ward V.J.
        • Quinn M.J.
        • Zidar F.J.
        • Plow E.F.
        • Topol E.J.
        • Byzova T.V.
        J. Biol. Chem. 2004; 279: 21651-21657
        • Zwicker J.I.
        • Peyvandi F.
        • Palla R.
        • Lombardi R.
        • Canciani M.T.
        • Cairo A.
        • Ardissino D.
        • Bernardinelli L.
        • Bauer K.A.
        • Lawler J.
        • Mannucci P.
        Blood. 2006; 108: 1280-1283
        • Miao W.M.
        • Seng W.L.
        • Duquette M.
        • Lawler P.
        • Laus C.
        • Lawler J.
        Cancer Res. 2001; 61: 7830-7839
      2. Calzada, M. J., Kuznetsova, S. A., Sipes, J. M., Rodrigues, R. G., Cashel, J. A., Annis, D. S., Mosher, D. F., and Roberts, D. D. (2008) Matrix Biology, in press

        • Topol E.J.
        • McCarthy J.
        • Gabriel S.
        • Moliterno D.J.
        • Rogers W.J.
        • Newby L.K.
        • Freedman M.
        • Metivier J.
        • Cannata R.
        • O'Donnell C.J.
        • Kottke-Marchant K.
        • Murugesan G.
        • Plow E.F.
        • Stenina O.
        • Daley G.Q.
        Circulation. 2001; 104: 2641-2644
        • Posey K.L.
        • Hayes E.
        • Haynes R.
        • Hecht J.T.
        Int. J. Biochem. Cell Biol. 2004; 36: 1005-1012
        • Hannah B.L.
        • Misenheimer T.M.
        • Annis D.S.
        • Mosher D.F.
        J. Biol. Chem. 2003; 278: 8929-8934
        • Hannah B.L.
        • Misenheimer T.M.
        • Pranghofer M.M.
        • Mosher D.F.
        J. Biol. Chem. 2004; 279: 51915-51922
        • Annis D.S.
        • Gunderson K.A.
        • Mosher D.F.
        J. Biol. Chem. 2007; 282: 27067-27075
        • Mosher D.F.
        • Huwiler K.G.
        • Misenheimer T.M.
        • Annis D.S.
        Methods Cell Biol. 2002; 69: 69-81
        • Mach H.
        • Middaugh C.R.
        • Lewis R.V.
        Anal. Biochem. 1992; 200: 74-80
        • Potterton L.
        • McNicholas S.
        • Krissinel E.
        • Gruber J.
        • Cowtan K.
        • Emsley P.
        • Murshudov G.N.
        • Cohen S.
        • Perrakis A.
        • Noble M.
        Acta Crystallogr. Sect. D Biol. Crystallogr. 2004; 60: 2288-2294
        • Tooney P.A.
        • Sakai T.
        • Sakai K.
        • Aeschlimann D.
        • Mosher D.F.
        Matrix Biol. 1998; 17: 131-143
      3. Otwinowski, Z., and Minor, W. (1997) in Methods in Enzymology (Carter, C. W., and Sweet, R. M., eds) Vol. 276 A, pp. 307–326, Academic Press, New York

        • Navaza J.
        Acta Crystallogr. Sect. D Biol. Crystallogr. 2001; 57: 1367-1372
        • CCP4
        Acta Crystallogr. Sect. D Biol. Crystallogr. 1994; 50: 760-763
        • Jones T.A.
        • Zou J.Y.
        • Cowan S.W.
        • Kjeldgaard M.
        Acta Crystallogr. Sect. A. 1991; 47: 110-119
        • Emsley P.
        • Cowtan K.
        Acta Crystallogr. Sect. D Biol. Crystallogr. 2004; 60: 2126-2132
        • Misenheimer T.M.
        • Hannah B.L.
        • Annis D.S.
        • Mosher D.F.
        Biochemistry. 2003; 42: 5125-5132
        • Misenheimer T.M.
        • Mosher D.F.
        J. Biol. Chem. 2005; 280: 41229-41235
        • Lawler J.
        • Sunday M.
        • Thibert V.
        • Duquette M.
        • George E.L.
        • Rayburn H.
        • Hynes R.
        J. Clin. Investig. 1998; 101: 982-992
        • Kyriakides T.R.
        • Zhu Y.-H.
        • Smith L.T.
        • Bain S.D.
        • Yang Z.
        • Lin M.T.
        • Danielson K.G.
        • Iozzo R.V.
        • LaMarca M.
        • McKinney C.E.
        • Ginns E.I.
        J. Cell Biol. 1998; 140: 419-430
        • Hankenson K.D.
        • Hormuzdi S.G.
        • Meganck J.A.
        • Bornstein P.
        Mol. Cell. Biol. 2005; 25: 5599-5606
        • Svensson L.
        • Aszodi A.
        • Heinegard D.
        • Hunziker E.B.
        • Reinholt F.P.
        • Fassler R.
        • Oldberg A.
        Mol. Cell. Biol. 2002; 22: 4366-4371
        • Breitwieser G.E.
        Int. J. Biochem. Cell Biol. 2008; 40: 1467-1480
        • Deere M.
        • Sanford T.
        • Francomano C.A.
        • Daniels K.
        • Hecht J.T.
        Am. J. Med. Genet. 1999; 85: 486-490
        • Boekholdt S.M.
        • Trip M.D.
        • Peters R.J.G.
        • Engelen M.
        • Boer J.M.A.
        • Feskens E.J.M.
        • Zwinderman A.H.
        • Kastelein J.J.P.
        • Reitsma P.H.
        Arterioscler. Thromb. Vasc. Biol. 2002; 22: e24-e27
        • Kleerekoper Q.
        • Hecht J.T.
        • Putkey J.A.
        J. Biol. Chem. 2002; 277: 10581-10589
        • Chen H.
        • Deere M.
        • Hecht J.T.
        • Lawler J.
        J. Biol. Chem. 2000; 275: 26538-26544
        • Hou J.
        • Putkey J.A.
        • Hecht J.T.
        Cell Calcium. 2000; 27: 309-314
        • Maddox B.K.
        • Mokashi A.
        • Keene D.R.
        • Bachinger H.P.
        J. Biol. Chem. 2000; 275: 11412-11417
        • Thur J.
        • Rosenberg K.
        • Nitsche D.P.
        • Pihlajamaa T.
        • Ala-Kokko L.
        • Heinegard D.
        • Paulsson M.
        • Maurer P.
        J. Biol. Chem. 2001; 276: 6083-6092
        • Kennedy J.
        • Jackson G.
        • Ramsden S.
        • Taylor J.
        • Newman W.
        • Wright M.J.
        • Donnai D.
        • Elles R.
        • Briggs M.D.
        Eur. J. Hum. Genet. 2005; 13: 547-555
        • DeLano W.L.
        PyMol. DeLano Scientific, San Carlos, CA2002