Published online before print March 27, 2008, doi: 10.1161/CIRCRESAHA.107.166124
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2008 American Heart Association, Inc.
Molecular Medicine |
Proline and 
-Carboxylated Glutamate Residues in Matrix Gla Protein Are Critical for Binding of Bone Morphogenetic Protein-4
From the Division of Cardiology (Y.Y., A.S., K.I.B.), David Geffen School of Medicine; and Molecular Biology Institute (K.I.B.), University of California, Los Angeles.
Correspondence to Kristina I. Boström, MD, PhD, Division of Cardiology, David Geffen School of Medicine at UCLA, Box 951679, Los Angeles, CA 90095-1679. E-mail kbostrom{at}mednet.ucla.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Arterial calcification is ubiquitous in vascular disease and is, in part, prevented by matrix Gla protein (MGP). MGP binds calcium ions through
-carboxylated glutamates (Gla residues) and inhibits bone morphogenetic protein (BMP)-2/-4. We hypothesized that a conserved proline (Pro)64 is essential for BMP inhibition. We further hypothesized that calcium binding by the Gla residues is a prerequisite for BMP inhibition. Site-directed mutagenesis was used to modify Pro64 and the Gla residues, and the effect on BMP-4 activity, and binding of BMP-4 and calcium was tested using luciferase reporter gene assays, coimmunoprecipitation, crosslinking, and calcium quantification. The results showed that Pro64 was critical for binding and inhibition of BMP-4 but not for calcium binding. The Gla residues were also required for BMP-4 binding but flexibility existed. As long as 1 Gla residue remained on each side of Pro64, the ability to bind and inhibit BMP-4 was preserved. Chelation of calcium ions by EDTA or warfarin treatment of cells led to loss of ability of MGP to bind BMP-4. Our results also showed that phenylalanine could replace Pro64 without loss of function and that zebrafish MGP, which lacks upstream Gla residues, did not function as a BMP inhibitor. The effect of MGP mutagenesis on vascular calcification was determined in calcifying vascular cells. Only MGP proteins with preserved ability to bind and inhibit BMP-4 prevented osteogenic differentiation and calcification. Together, our results suggest that BMP and calcium binding in MGP are independent but functionally intertwined processes and that the BMP binding is essential for prevention of vascular calcification.
Key Words: matrix GLA protein • bone morphogenetic protein • vascular calcification • BMP inhibitor • mutagenesis
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Matrix Gla protein (MGP) has been identified as an inhibitor of arterial calcification; gene deletion in mice leads to massive arterial calcification and early death from vascular rupture.1 Patients with defined mutations in the MGP gene (Keutel syndrome) have widespread calcification in the tracheobronchial tree, and the quality and duration of life appears to be determined by the extent of pulmonary disease.2,3 We and others have identified MGP as an inhibitor of bone morphogenetic protein (BMP).4–7 Specifically, we have shown that MGP binds and inhibits BMP-2 and -4, which are both detected in calcified atherosclerotic lesions.8–11 We have also shown that MGP-deficient cells have increased susceptibility to BMP-2–induced cell differentiation12 and that MGP interferes with BMP-4 activity in the developing lungs of MGP transgenic mice.13
MGP is known to be a calcium-binding protein. Calcium is bound through -carboxylated glutamates, so-called "Gla residues," and induces conformational changes in the MGP protein.4,14 Warfarin interferes with the vitamin K–dependent -carboxylation and renders MGP nonfunctional,15 presumably through loss of calcium binding. The majority of the Gla residues are located in the center of the MGP protein (amino acid residue 56 to 71 in human MGP), a region that has been implicated in the binding of BMP.4 However, the precise amino acid residues involved in BMP binding have not been identified.
In the BMP type I receptor, a phenylalanine residue (Phe85) plays a key role in binding BMP-2.16 The aromatic side chain of Phe85, with "knob-into-hole" packing, points into a hydrophobic pocket formed at the interfaces of 2 BMP-2 monomers. Similarly, a critical proline residue (Pro35) in the BMP inhibitor Noggin occupies the same pocket,17,18 thereby preventing ligand–receptor binding. When examining the protein sequence of human MGP, we found a conserved proline residue (Pro64) centrally located in the BMP-binding region, with 2 Gla residues on each side. This proline residue is conserved in MGP from all species so far examined (human, bovine, mouse, rat, chicken, Xenopus, and fish species including shark and zebrafish) (Figure I in the online data supplement). We hypothesized that it has a similar role to that of Pro35 in Noggin and that mutagenesis of this proline residue would abolish the ability of MGP to bind and inhibit BMP. We further hypothesized that removal of the Gla residues would affect the ability of MGP to bind and inhibit BMP.
In this study, we show that Pro64 is critical for binding and inhibition of BMP-4 but not for binding of calcium. Deletion of 2 Gla residue on 1 or the other side of Pro64 abolishes the ability to bind and inhibit BMP-4. Removal of calcium ions leads to loss of the ability of MGP to bind BMP-4, suggesting that calcium is essential for the use of Pro64 in binding BMP-4. Only MGP proteins with preserved ability to bind BMP-4 prevent mineralization of calcifying vascular cells. Together, our results suggest that BMP- and calcium binding in MGP are independent but functionally intertwined processes and that the BMP binding is essential for prevention of vascular calcification.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
An expanded Materials and Methods section is available in the online data supplement at http://circres.ahajournals.org.
Cell Culture and Transfection Assays
Bovine aortic endothelial cells (BAECs) and calcifying vascular cells (CVCs) were cultured and transfected as previous described.7,19–21 Luciferase assays were performed as previously described19 and normalized to Renilla. HEK293 and C2C12 cells were obtained from the American Type Culture Collection, and HEK293 cells were transfected using Superfect. Alkaline phosphatase activity and calcium deposition in CVCs were determined as previously described,20,22 after 2 and 8 days of BMP-4 treatment, respectively.
Vector Constructions and Mutagenesis
The BMP-responsive luciferase reporter gene (BRE-Luc) was obtained from Dr Peter ten Dijke (Leiden University Medical Center, Leiden, The Netherlands).23 The pN-FLAG-hMGP vector has been described,7 and the FLAG-tagged zebrafish MGP vector was constructed as described in the online data supplement. Mutagenesis was performed with the QuickSite mutagenesis kit using the primers in supplemental Table I and pN-FLAG-hMGP as template.
Immunoblotting, Immunoprecipitation, and Crosslinking
Immunoblotting and coimmunoprecipitation of BMP-4 and N-FLAG-MGP were performed as previously described.7,19 Serum-free conditioned media from transfected HEK293 cells were used in the crosslinking experiments. Conditioned medium was mixed with BMP-4 and chemically crosslinked with disuccinimidyl suberate as described previously.24 The cross-linked products were analyzed by immunoblotting or protein staining.
Quin2 Staining and Calcium Quantification
Media containing similar concentrations of proteins were concentrated, electrophoresed, and transferred to Hyperbond membranes. Quin2 staining was performed as previously described.25 Protein bands were excised and digested, and quantification of calcium was performed as previously described.20
Statistics
Data were analyzed for statistical significance by ANOVA with post hoc Scheffe’s analysis. The analyses were performed using StatView, version 4.51 (Abacus Concepts). All experiments were repeated a minimum of 3 times.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Mutagenesis of Pro64 Abolished the Ability of MGP to Bind and Inhibit BMP-4
To determine whether Pro64 was required for binding and inhibition of BMP-4, we mutated Pro64 in the N-FLAG-MGP expression vector to glycine (N-FLAG-MGP-P64G) using site-directed mutagenesis. We have shown previously that FLAG-tagged MGP is functionally equivalent to nontagged MGP.6,7 To compare the effects of N-FLAG-MGP and -MGP-P64G on BMP-4 activity, we cotransfected BAECs with a BMP-responsive reporter gene (BRE-Luc), a Renilla luciferase vector, and increasing amounts of expression constructs for N-FLAG-MGP or -MGP-P64G. The cells were treated with BMP-4 (40 ng/mL) for 24 hours, and luciferase activity was determined and normalized to Renilla luciferase. The results showed that N-FLAG-MGP progressively inhibited luciferase activity induced by BMP-4, whereas N-FLAG-MGP-P64G had no effect on the BMP activity (Figure 1A). The experiments were also performed in human aortic endothelial cells with similar results (data not shown). To ensure that the MGP proteins were secreted and to determine whether they interacted with BMP-4, we analyzed the media from BAECs transfected with the maximal amount of MGP construct. First, equal amounts of the media were analyzed by immunoblotting using anti-FLAG-antibodies. The results showed that the level of secreted N-FLAG-MGP and -MGP-P64G were similar (Figure 1B, top immunoblot). Second, BMP-4 was added to the N-FLAG-MGP and -MGP-P64G media, and coimmunoprecipitations were performed using anti-BMP-4 or anti-FLAG antibodies. The anti-BMP-4 immunoprecipitates were analyzed by immunoblotting using anti-FLAG-antibodies. Conversely, anti-FLAG immunoprecipitates were analyzed by immunoblotting using anti–BMP-4 antibodies. The results showed that coimmunoprecipitation of N-FLAG-MGP and BMP-4 occurred whether the complex was immunoprecipitated with antibodies to BMP-4 or the FLAG-tag (Figure 1B). However, no coimmunoprecipitation was detected for N-FLAG-MGP-P64G and BMP-4 (Figure 1B).
|
To further validate our results, we performed crosslinking experiments using BMP-4 and N-FLAG-MGP or -MGP-P64G. Serum-free conditioned media containing the respective MGP proteins were collected 24 hours after transfection of HEK293 cells. BMP-4 (6 to 10 ng) was added to 80 µL of conditioned medium and cross-linked with disuccinimidyl suberate, and protein complexes were analyzed by immunoblotting with anti-FLAG antibodies. Incubation of BMP-4 and N-FLAG-MGP led to the formation of a complex migrating with molecular mass of approximately 60 to 62 kDa that would correspond to a BMP-4 dimer (2x20 to 22 kDa) bound to a MGP dimer (2x10 to 11 kDa) (Figure 1C). By contrast, no complex formation was detected when N-FLAG-MGP-P64G was used. Complexes were also immunoprecipitated from large volumes of cross-linked material containing BMP-4 and N-FLAG-MGP or -MGP-P64G using anti-FLAG antibodies and analyzed by SDS-PAGE and protein staining. Again, the major complex was 60 to 62 kDa in samples containing N-FLAG-MGP, but small amounts of complexes of lower molecular masses were also detected (Figure 1D). Finally, the protein complexes were analyzed by immunoblotting using anti–BMP-4 antibodies. For these antibodies to work, 50 to 100 ng of BMP-4 was added to 80 µL of conditioned medium and cross-linked with disuccinimidyl suberate. The same 60- to 62-kDa complex previously seen with anti-FLAG antibodies and protein staining was detected in samples with N-FLAG-MGP but not with N-FLAG-MGP-P64G. In addition, 2 of the complexes detected by protein staining reacted with anti–BMP-4. Most likely, these corresponded to a BMP-4 monomer bound to either an MGP monomer or an MGP dimer (Figure 1E). No complex formation was detected with N-FLAG-MGP-P64G.
Together, the crosslinking results were consistent with the coimmunoprecipitations and suggested that the ability to bind BMP-4 was lost in N-FLAG-MGP-P64G. Furthermore, the results suggested that a dimer of BMP-4 binds to a dimer of MGP.
To ensure that loss of MGP function was not unique to the replacement of proline with glycine, we mutated Pro64 to alanine or tyrosine. Both were unable to replace proline and preserve MGP function (see the online data supplement). Furthermore, mutagenesis of lysine-63 and valine-65, which surround Pro64, or insertion of a serine between lysine-63 and Pro64, did not affect MGP function. The results showed that only the mutants with intact Pro64 bound and inhibited BMP-4, which further supports that Pro64 is essential for MGP function (see online data supplement).
Mutagenesis of Gla Residues Differentially Affects the Ability of MGP to Bind and Inhibit BMP-4
To determine whether the Gla residues are essential for binding and inhibition of BMP-4, we modified the Gla residues by mutating glutamate (Glu)56, -60, -67, and -71 to glycines in the N-FLAG-MGP expression construct using site-directed mutagenesis. We constructed 10 MGP vectors with different combinations of Glu mutations (shown in Figure 2A through 2C). Three combinations involved the glutamates upstream of Pro64 (Figure 2A), 3 involved the glutamates downstream of Pro64 (Figure 2B), and 4 involved glutamates on both sides of Pro64 (Figure 2C).
|
We determined the effect of the mutated MGP proteins on BMP-4 activity in BAECs by luciferase reporter gene assays and compared them with nonmutated MGP. For the MGP proteins with mutations of the upstream glutamates, mutagenesis of either Glu56 or Glu60 did not diminish the ability of MGP to inhibit BMP-4, but mutagenesis of both Glu56 and Glu60 abolished BMP-4 inhibition (Figure 2A, left). Similar results were obtained for the MGP proteins with mutations of the downstream glutamates. Mutagenesis of either Glu67 or Glu71 did not diminish BMP inhibition, but mutagenesis of both Glu67 and Glu71 abolished BMP inhibition (Figure 2B, left). We then generated 4 constructs with 1 mutated glutamate on each side of Pro64. The results showed that all mutated MGP proteins retained the ability to inhibit BMP even though there may be differences in the effectiveness of the inhibition (Figure 2C, left).
Analysis of the media from the transfected cells showed that the levels of secreted mutated MGP proteins were similar to that of nonmutated MGP (Figure 2A through 2C, right, top anti-FLAG immunoblots). To assess protein interactions, BMP-4 was added to the respective media, and coimmunoprecipitations were performed and analyzed as described above. Overall, the results from the coimmunoprecipitations showed that interaction with BMP-4 was abolished for the same mutated MGP proteins that had lost the ability to inhibit BMP-4. MGP proteins with 2 Gla residues removed on the same side of Pro64 were no longer able to bind BMP-4 (Figure 2A through 2B, left graphs, PL3 and PR3). MGP proteins with only 1 Glu mutation, or with 1 Glu mutation on each side of Pro64, retained BMP-4 binding (Figure 2A through 2C). The results suggest that a minimum of 1 Gla residue on each side of Pro64 is required for BMP-4 inhibition and binding.
Mutagenesis of Pro64 Does Not Affect Calcium Binding by MGP
Other investigators have shown that MGP undergoes conformational changes when it binds calcium.4,14 To determine whether mutagenesis of Pro64 to glycine affected calcium binding, N-FLAG-MGP and -MGP-P64G were analyzed using quin2 staining and subsequent quantification of bound calcium. Samples containing similar concentrations of the respective MGP proteins were electrophoresed and transferred to Hyperbond membranes. Troponin C and transforming growth factor-β1 were included as a calcium-binding and a non–calcium-binding control protein, respectively. The MGP protein bands were visualized using quin2, the bands were excised, and the calcium was released and quantified. The results showed no significant difference in calcium content for N-FLAG-MGP and -MGP-P64G (Figure 3A). Low amounts of calcium were detected also after transfection with empty plasmid and were likely derived from endogenous MGP secretion.
|
To determine whether BMP binding is abolished by calcium removal, N-FLAG-MGP was treated with increasing concentrations of EDTA and then coimmunoprecipitated with BMP-4 as described above. Antigen binding by the antibodies used for immunoprecipitation was not sensitive to calcium concentrations. The results showed that calcium depletion abolished the coimmunoprecipitation of BMP-4 and N-FLAG-MGP (Figure 3B, right 3 lanes), providing further support that calcium binding in MGP is essential for BMP binding.
Calcium binding is also prevented by mutating all 4 Gla residues or treating the cells with warfarin, which inhibits the vitamin K–dependent -carboxylation. We determined the effect of both interventions on the ability of MGP to bind and inhibit BMP-4. We used an expression construct where all 4 glutamates surrounding Pro64 were mutated to glycines (N-FLAG-MGP-4GlaG), or we treated BAECs with warfarin (15 µg/mL) after transfection of the N-FLAG-MGP construct. The effect of the MGP proteins on BMP-4 activity was tested as before by cotransfection of the BMP-responsive luciferase reporter gene, followed by treatment with BMP-4 (40 ng/mL). The results showed that neither N-FLAG-4GlaG nor MGP from warfarin-treated cells inhibited BMP-4 signaling activity (Figure 3C). Analysis of the media from the transfected and warfarin-treated cells showed that the MGP proteins were similarly expressed (Figure 3D, top anti-FLAG immunoblot). However, N-FLAG-MGP-4GlaG and N-FLAG-MGP from warfarin-treated cells did not interact with BMP-4, as determined by coimmunoprecipitation (Figure 3D), further supporting that calcium binding is critical for BMP binding and BMP inhibition.
Replacement of Pro64 With Phenylalanine Preserves the Ability of MGP to Bind and Inhibit BMP-4
Our hypothesis predicted that replacement of Pro64 by phenylalanine (F) would preserve the MGP function because phenylalanine is the key residue in the binding of BMP-2 by the BMP type I receptor.16 To test this, we mutated N-FLAG-MGP to -MGP-P64F and then determined the effect of N-FLAG-MGP-P64F on BMP-4 activity by luciferase reporter gene assays. The results showed that N-FLAG-MGP-P64F inhibited BMP-4 activity at least as well as N-FLAG-MGP (Figure 4A, left).
|
Analysis of the media showed that the expression level of N-FLAG-MGP-P64F was similar to that of N-FLAG-MGP (Figure 4A, right, top anti-FLAG immunoblot). BMP-4 was added to the N-FLAG-MGP and -MGP-P64F media, and protein interactions were examined by coimmunoprecipitations. The results showed that coimmunoprecipitation was similar for N-FLAG-MGP and -MGP-P64F (Figure 4A, right), suggesting that proline and phenylalanine are interchangeable in the 64 position.
Zebrafish MGP Does Not Affect BMP-4
Fish MGP lacks the 2 glutamates upstream of the conserved proline (supplemental Figure I), which suggests that fish MGP does not function as a BMP inhibitor. To determine whether fish MGP binds and inhibits BMP-4 in our system, we constructed expression vectors for untagged and N-terminally FLAG-tagged zebrafish MGP. The effect of the zebrafish MGP on BMP-4 signaling activity was determined using luciferase reporter gene assays in BAECs and compared to human MGP. The results showed that untagged and FLAG-tagged zebrafish MGP had no effect on BMP-4 signaling (Figure 4B, left). Analysis of the media showed a similar expression level of the 2 MGP proteins (Figure 4B, right, top anti-FLAG immunoblot). BMP-4 was then added to the respective medium and coimmunoprecipitations were performed. The results showed no coimmunoprecipitation of BMP-4 and N-FLAG zebrafish MGP (Figure 4B, right) suggesting that fish MGP is not a BMP inhibitor.
MGP With Lost Ability to Bind and Inhibit BMP-4 Does Not Prevent Mineralization in Calcifying Vascular Cells
MGP is known as an inhibitor of vascular calcification. To determine whether mutant MGP that has lost the ability to bind BMP-4 still inhibits calcification, we used the calcifying CVCs, a well-established model of vascular calcification that mineralizes in response to BMP.20,26 The CVCs were transiently transfected with empty vector (EV) or expression constructs for wild-type or mutant N-FLAG-MGP proteins. Expression of the MGP proteins was verified by immunoblotting of culture media collected 48 hours after transfection using anti-FLAG antibodies. The expression of the different MGP proteins was similar (Figure 5C and 5D). The CVCs were treated for up to 8 days with BMP-4 (20 ng/mL). The activity of alkaline phosphatase, an early marker of osteogenic differentiation, and calcium mineral deposition were determined after 2 and 8 days, respectively. The results showed that both alkaline phosphatase and mineral deposition were only inhibited in CVCs transfected with MGP proteins known to bind and inhibit BMP-4 (Figure 5C and 5D). The results suggest that binding of BMP-4, but not of calcium, is essential for the prevention of vascular calcification by MGP.
|
Similarly, we also determined that N-FLAG-MGP-P64G was unable to prevent BMP-induced osteogenic differentiation in the pluripotent C2C12 cells, as determined by morphology and alkaline phosphate activity (see online data supplement).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MGP is best known as an inhibitor of vascular calcification and may interfere with BMP-signaling during the process of calcification. Others and we have demonstrated that MGP binds and inhibits BMP-2 and -4 in vascular cells.4–7 We have also shown that MGP-deficient cells have an increased susceptibility for BMP-2–induced cell differentiation12,20 and that excess MGP interferes with BMP-4 activity in the developing lungs of MGP transgenic mice.13 In this study, we show that Pro64 in MGP is critical for binding of BMP-4 but not of calcium. Deletion of 2 Gla residues on 1 or the other side of Pro64 also results in lost inhibition and binding of BMP-4, suggesting that the Gla residues contribute to the conformation required for BMP binding.
Proline and phenylalanine have been shown to be critical for BMP binding by Noggin and the BMP type I receptor, respectively.16–18 Our results show that MGP binds and inhibits BMP-4 with either proline or phenylalanine in the 64 position, suggesting that the binding between MGP and BMP-4 occurs according to the same knob-into-hole model that has been proposed for Noggin and the BMP type I receptor based on crystal structures.16,18
Differences in activity between different MGP mutants were noted in the BMP-4 activity curves (see, eg, Figure 2). It is possible that these curves reflect the fit between MGP and BMP-4, that a stronger inhibition of BMP-4 may indicate a better fit and vice versa. The relative ability of a specific MGP mutant to inhibit BMP-4, as compared with other mutants, was consistent among experiments, although the absolute differences varied. However, the immunoprecipitation experiments did not detect any differences in the ability to bind BMP-4, which may be attributable to technical factors. The MGP media used for immunoprecipitations were all collected from cells transfected with the maximal amount of MGP construct, but BMP-4 was still added in excess for the BMP-4 antibodies to work optimally. Thus, even if the affinity for BMP-4 of some MGP mutants were low, it may not be detected because of the excess BMP-4. In addition, the immunoblotting was technically difficult to perform because of the proximity of the light chains (from precipitating antibodies), which may have affected transfer and detection of BMP-4 and MGP. Further experimentation will be needed to clarify the "molecular fit" between BMP-4 and MGP.
Overall, the crosslinking data support the formation of a complex containing a BMP-4 dimer and an MGP dimer. Small amounts of presumed intermediary complexes containing a BMP-4 monomer with 1 or 2 MGP were also detected when the immunoblots were analyzed with anti-BMP-4 antibodies (Figure 1D).
Our data suggest that the Gla residues are intimately involved in achieving the conformation necessary for BMP binding, which has previously been suggested by the results of Sweatt et al5 that show that BMP-2 binds to the Gla-containing regions of MGP only in the presence of calcium. Similarly, the Gla residues have been implicated in structure–function in other proteins. In protein C, an anticoagulant, and factor VII, a procoagulant, the Gla residues have been implicated in membrane association.27 In osteocalcin, a calcium-binding protein of bone, 3 Gla residues are part of a helical motif and participates directly in the binding of calcium and the adsorption of osteocalcin to hydroxylapatite surfaces.28 The spacing of the Gla residues in osteocalcin (at positions 17, 21, and 24) places all Gla side chains on the same face of 1 
helix, closely paralleling the interatomic separation of calcium in the hydroxylapatite lattice. Interestingly, the spacing of MGP Gla residues 56 and 60, and 67 and 71, respectively, is identical to the spacing of the osteocalcin Gla residues, suggesting that the Gla residues in MGP are part of helical motifs on both sides of Pro64.
The mechanism of MGP in vascular calcification continues to be under debate. On one hand, it has been proposed that MGP inhibits vascular calcification by binding calcium ions, thereby preventing mineral precipitation or crystal growth.29 On the other hand, others and we have proposed that MGP inhibits BMP-signaling, thereby preventing unwanted cell-induced mineralization.4,6 The early and profound changes of smooth muscle cells into chondrocyte-like cells reported for the MGP knockout mice1,30,31 are consistent with increased susceptibility to BMP in the MGP-deficient cells. Our previous studies also showed an increased susceptibility to BMP-2–induced cell differentiation in vitro,12 which is further supported by the findings in this study. Our results showed that only those MGP proteins with the ability to bind BMP-4 inhibited mineralization (Figure 5). Specifically, the mutant N-FLAG-MGP-P64G, which was shown to bind calcium (Figure 3A), did not prevent calcium deposition in CVCs. Thus, our results support the hypothesis that MGP prevents vascular calcification by inhibiting BMP rather than by binding calcium. However, our results suggest that calcium is intimately involved in enabling MGP to inhibit BMP by binding to the Gla residues. We speculate that the MGP may have different conformational states. On one hand, Gla residues may bind to membranes or mineral crystals and possibly preclude BMP binding. On the other hand, binding of soluble calcium may induce structural transitions that allow BMP binding.
Although MGP is similar to Noggin in that both may be considered morphogens and both has proline residues that are critical for BMP binding, MGP differs in its dependency on calcium for BMP binding. Because MGP is able to interact with vitronectin in extracellular matrix,32 it is possible that MGP is more aimed at regulating local BMP and may be more sensitive to variations in calcium concentrations. MGP may be released by cleavage of the C terminus, which has been shown to be missing in MGP purified from tissue (see Figure 6 for hypothetical model).33 Noggin, on the other hand, may have a larger range of action and be independent of calcium concentrations.
|
Our results may also relate to the pathogenesis of a debilitating form of arteriolar calcification, so-called calcific uremic arteriolopathy, or calciphylaxis, which may affect dialysis patients.34 BMP-4 has been identified in the perivascular areas of calciphylaxis lesions,35 and warfarin has been proposed to be a risk factor for calciphylaxis.34 Pericytes, which are perivascular cells in small vessels, are similar to CVCs8 and are known to express MGP.36 It is possible that a loss of functional MGP, eg, after warfarin treatment, triggers a BMP-4–induced process that leads to calciphylaxis.
Our results may have implications for MGP function in different species. Mouse MGP lacks the first -carboxylated Glu residue in the BMP-binding region when compared with human MGP. However, mutagenesis of this Glu residue did not diminish inhibition or binding of BMP-4, and we would therefore expect MGP to function similarly in mouse and human. This is supported by the similarities in phenotype between the MGP knockout mouse and patients with the Keutel syndrome, which include underdevelopment of facial bones, short stature, and abnormal cartilage calcification.1,3 On the other hand, fish MGP lacks both Gla residues upstream of Pro64, and our results suggest that zebrafish MGP does not inhibit BMP. It is still possible that it would bind zebrafish BMP-4 even though there is a very high degree of homology between human and zebrafish BMP-4. It is intriguing that MGP from species with lungs all have conserved glutamates upstream of the conserved proline (supplemental Figure I), whereas none of the fish species do. It is possible that MGP took on the role as a BMP inhibitor as lungs were developed. We have previously shown that MGP has a regulatory effect on vascular development in the lungs13 that contributes to the coordination of airway and vascular development.
In summary, our study shows that the presence of Pro64 and a sufficient number of Gla residues in MGP are essential for binding of BMP-4 and suggests that BMP- and calcium binding in MGP are independent but functionally intertwined processes. Our study also shows that the BMP binding is essential for prevention of vascular calcification by MGP.
|
Acknowledgments |
|---|
Sources of Funding
This work was funded in part by NIH grants HL30568 and HL81397, the American Heart Association (Western Affiliate), and the International HDL Awards Program (Pfizer).
Disclosures
None.
|
Footnotes |
|---|
This manuscript was sent to Mark Taubman, Consulting Editor, for review by expert referees, editorial decision, and final disposition.
Original received October 16, 2007; revision received March 12, 2008; accepted March 13, 2008.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1. Luo G, Ducy P, McKee MD, Pinero GJ, Loyer E, Behringer RR, Karsenty G. Spontaneous calcification of arteries and cartilage in mice lacking matrix GLA protein. Nature. 1997; 386: 78–81.[CrossRef][Medline] [Order article via Infotrieve]
2. Munroe PB, Olgunturk RO, Fryns JP, Van Maldergem L, Ziereisen F, Yuksel B, Gardiner RM, Chung E. Mutations in the gene encoding the human matrix Gla protein cause Keutel syndrome. Nat Genet. 1999; 21: 142–144.[CrossRef][Medline] [Order article via Infotrieve]
3. Hur DJ, Raymond GV, Kahler SG, Riegert-Johnson DL, Cohen BA, Boyadjiev SA. A novel MGP mutation in a consanguineous family: review of the clinical and molecular characteristics of Keutel syndrome. Am J Med Genet A. 2005; 135: 36–40.[Medline] [Order article via Infotrieve]
4. Wallin R, Cain D, Hutson SM, Sane DC, Loeser R. Modulation of the binding of matrix Gla protein (MGP) to bone morphogenetic protein-2 (BMP-2). Thromb Haemost. 2000; 84: 1039–1044.[Medline] [Order article via Infotrieve]
5. Sweatt A, Sane DC, Hutson SM, Wallin R. Matrix Gla protein (MGP) and bone morphogenetic protein-2 in aortic calcified lesions of aging rats. J Thromb Haemost. 2003; 1: 178–185.[CrossRef][Medline] [Order article via Infotrieve]
6. Zebboudj AF, Imura M, Bostrom K. Matrix GLA protein, a regulatory protein for bone morphogenetic protein-2. J Biol Chem. 2002; 277: 4388–4394.
7. Yao Y, Zebboudj AF, Shao E, Perez M, Bostrom K. Regulation of bone morphogenetic protein-4 by matrix GLA protein in vascular endothelial cells involves activin-like kinase receptor 1. J Biol Chem. 2006; 281: 33921–33930.
8. Bostrom K, Watson KE, Horn S, Wortham C, Herman IM, Demer LL. Bone morphogenetic protein expression in human atherosclerotic lesions. J Clin Invest. 1993; 91: 1800–1809.[Medline] [Order article via Infotrieve]
9. Dhore CR, Cleutjens JP, Lutgens E, Cleutjens KB, Geusens PP, Kitslaar PJ, Tordoir JH, Spronk HM, Vermeer C, Daemen MJ. Differential expression of bone matrix regulatory proteins in human atherosclerotic plaques. Arterioscler Thromb Vasc Biol. 2001; 21: 1998–2003.
10. Schluesener HJ, Meyermann R. Immunolocalization of BMP-6, a novel TGF-beta-related cytokine, in normal and atherosclerotic smooth muscle cells. Atherosclerosis. 1995; 113: 153–156.[CrossRef][Medline] [Order article via Infotrieve]
11. Davies MR, Lund RJ, Hruska KA. BMP-7 is an efficacious treatment of vascular calcification in a murine model of atherosclerosis and chronic renal failure. J Am Soc Nephrol. 2003; 14: 1559–1567.
12. Bostrom K, Tsao D, Shen S, Wang Y, Demer LL. Matrix GLA protein modulates differentiation induced by bone morphogenetic protein-2 in C3H10T1/2 cells. J Biol Chem. 2001; 276: 14044–14052.
13. Yao Y, Nowak S, Yochelis A, Garfinkel A, Bostrom KI. Matrix GLA protein, an inhibitory morphogen in pulmonary vascular development. J Biol Chem. 2007; 282: 30131–30142.
14. Hackeng TM, Rosing J, Spronk HM, Vermeer C. Total chemical synthesis of human matrix Gla protein. Protein Sci. 2001; 10: 864–870.[CrossRef][Medline] [Order article via Infotrieve]
15. Price PA, Faus SA, Williamson MK. Warfarin causes rapid calcification of the elastic lamellae in rat arteries and heart valves. Arterioscler Thromb Vasc Biol. 1998; 18: 1400–1407.
16. Kirsch T, Sebald W, Dreyer MK. Crystal structure of the BMP-2-BRIA ectodomain complex. Nat Struct Biol. 2000; 7: 492–496.[CrossRef][Medline] [Order article via Infotrieve]
17. Groppe J, Greenwald J, Wiater E, Rodriguez-Leon J, Economides AN, Kwiatkowski W, Affolter M, Vale WW, Belmonte JC, Choe S. Structural basis of BMP signalling inhibition by the cystine knot protein Noggin. Nature. 2002; 420: 636–642.[CrossRef][Medline] [Order article via Infotrieve]
18. Shi Y, Massague J. Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell. 2003; 113: 685–700.[CrossRef][Medline] [Order article via Infotrieve]
19. Bostrom K, Zebboudj AF, Yao Y, Lin TS, Torres A. Matrix GLA protein stimulates VEGF expression through increased transforming growth factor-beta1 activity in endothelial cells. J Biol Chem. 2004; 279: 52904–52913.
20. Zebboudj AF, Shin V, Bostrom K. Matrix GLA protein and BMP-2 regulate osteoinduction in calcifying vascular cells. J Cell Biochem. 2003; 90: 756–765.[CrossRef][Medline] [Order article via Infotrieve]
21. Yao Y, Zebboudj AF, Torres A, Shao E, Bostrom K. Activin-like kinase receptor 1 (ALK1) in atherosclerotic lesions and vascular mesenchymal cells. Cardiovasc Res. 2007; 74: 279–289.
22. Parhami F, Morrow AD, Balucan J, Leitinger N, Watson AD, Tintut Y, Berliner JA, Demer LL. Lipid oxidation products have opposite effects on calcifying vascular cell and bone cell differentiation. A possible explanation for the paradox of arterial calcification in osteoporotic patients. Arterioscler Thromb Vasc Biol. 1997; 17: 680–687.
23. Goumans MJ, Valdimarsdottir G, Itoh S, Lebrin F, Larsson J, Mummery C, Karlsson S, ten Dijke P. Activin receptor-like kinase (ALK)1 is an antagonistic mediator of lateral TGFbeta/ALK5 signaling. Mol Cell. 2003; 12: 817–828.[CrossRef][Medline] [Order article via Infotrieve]
24. Piccolo S, Sasai Y, Lu B, De Robertis EM. Dorsoventral patterning in Xenopus: inhibition of ventral signals by direct binding of chordin to BMP-4. Cell. 1996; 86: 589–598.[CrossRef][Medline] [Order article via Infotrieve]
25. Tatsumi R, Shimada K, Hattori A. Fluorescence detection of calcium-binding proteins with quinoline Ca-indicator quin2. Anal Biochem. 1997; 254: 126–131.[CrossRef][Medline] [Order article via Infotrieve]
26. Watson KE, Bostrom K, Ravindranath R, Lam T, Norton B, Demer LL. TGF-beta 1 and 25-hydroxycholesterol stimulate osteoblast-like vascular cells to calcify. J Clin Invest. 1994; 93: 2106–2113.[Medline] [Order article via Infotrieve]
27. Nelsestuen GL. Enhancement of vitamin-K-dependent protein function by modification of the gamma-carboxyglutamic acid domain: studies of protein C and factor VII. Trends Cardiovasc Med. 1999; 9: 162–167.[CrossRef][Medline] [Order article via Infotrieve]
28. Hauschka PV. Osteocalcin: the vitamin K-dependent Ca2+-binding protein of bone matrix. Haemostasis. 1986; 16: 258–272.[Medline] [Order article via Infotrieve]
29. Schinke T, McKee MD, Karsenty G. Extracellular matrix calcification: where is the action? Nat Genet. 1999; 21: 150–151.[CrossRef][Medline] [Order article via Infotrieve]
30. El-Maadawy S, Kaartinen MT, Schinke T, Murshed M, Karsenty G, McKee MD. Cartilage formation and calcification in arteries of mice lacking matrix Gla protein. Connect Tissue Res. 2003; 44 (Suppl 1): 272–278.[CrossRef][Medline] [Order article via Infotrieve]
31. Steitz SA, Speer MY, Curinga G, Yang HY, Haynes P, Aebersold R, Schinke T, Karsenty G, Giachelli CM. Smooth muscle cell phenotypic transition associated with calcification: upregulation of Cbfa1 and downregulation of smooth muscle lineage markers. Circ Res. 2001; 89: 1147–1154.
32. Nishimoto SK, Nishimoto M. Matrix Gla protein C-terminal region binds to vitronectin. Co-localization suggests binding occurs during tissue development. Matrix Biol. 2005; 24: 353–361.[CrossRef][Medline] [Order article via Infotrieve]
33. Hale JE, Williamson MK, Price PA. Carboxyl-terminal proteolytic processing of matrix Gla protein. J Biol Chem. 1991; 266: 21145–21149.
34. Coates T, Kirkland GS, Dymock RB, Murphy BF, Brealey JK, Mathew TH, Disney AP. Cutaneous necrosis from calcific uremic arteriolopathy. Am J Kidney Dis. 1998; 32: 384–391.[Medline] [Order article via Infotrieve]
35. Griethe W, Schmitt R, Jurgensen JS, Bachmann S, Eckardt KU, Schindler R. Bone morphogenic protein-4 expression in vascular lesions of calciphylaxis. J Nephrol. 2003; 16: 728–732.[Medline] [Order article via Infotrieve]
36. Canfield AE, Doherty MJ, Kelly V, Newman B, Farrington C, Grant ME, Boot-Handford RP. Matrix Gla protein is differentially expressed during the deposition of a calcified matrix by vascular pericytes. FEBS Lett. 2000; 487: 267–271.[CrossRef][Medline] [Order article via Infotrieve]
This article has been cited by other articles:
 |
|
 |
|
  D. Umulis, M. B. O'Connor, and S. S. Blair The extracellular regulation of bone morphogenetic protein signaling Development, November 15, 2009; 136(22): 3715 - 3728. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C McCann and B. N Ames Vitamin K, an example of triage theory: is micronutrient inadequacy linked to diseases of aging? Am. J. Clinical Nutrition, October 1, 2009; 90(4): 889 - 907. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Yao, A. D. Watson, S. Ji, and K. I. Bostrom Heat Shock Protein 70 Enhances Vascular Bone Morphogenetic Protein-4 Signaling by Binding Matrix Gla Protein Circ. Res., September 11, 2009; 105(6): 575 - 584. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. S. Shao, L. Lin, Y. Yao, and K. I. Bostrom Expression of vascular endothelial growth factor is coordinately regulated by the activin-like kinase receptors 1 and 5 in endothelial cells Blood, September 3, 2009; 114(10): 2197 - 2206. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Mizobuchi, D. Towler, and E. Slatopolsky Vascular Calcification: The Killer of Patients with Chronic Kidney Disease J. Am. Soc. Nephrol., July 1, 2009; 20(7): 1453 - 1464. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Yao, E. S. Shao, M. Jumabay, A. Shahbazian, S. Ji, and K. I. Bostrom High-Density Lipoproteins Affect Endothelial BMP-Signaling by Modulating Expression of the Activin-Like Kinase Receptor 1 and 2 Arterioscler Thromb Vasc Biol, December 1, 2008; 28(12): 2266 - 2274. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||