Harvey A. Goldberg

Osteopontin modulates CD44-dependent chemotaxis of peritoneal macrophages through G-protein-coupled receptors: evidence of a role for an intracellular form of osteopontin.

Expression of osteopontin (OPN) by activated T‐cells and macrophages is required for the development of cell‐mediated inflammatory responses. Acting through integrin αvβ3 and CD44 receptors, OPN can promote chemoattraction and pro‐inflammatory cytokine expression by macrophages. In this study, we have used periotoneal macrophages from OPN−/, CD44−/−, and WT mice to study the relationship between OPN and CD44 in macrophage migration. Using confocal microscopy, we show that OPN co‐distributes with CD44 inside macrophages at cell edges and in cell processes in a mutually dependent manner. The existence of an intracellular form of OPN is supported by pulse‐chase studies in which a thrombin‐sensitive, phosphorylated protein immunoprecipitated with OPN antibodies is retained inside macrophages. In OPN−/− and CD44−/− macrophages, the absence of CD44 and OPN, respectively, is associated with the formation of fewer cell processes, reduced cell fusion required to form functional multinucleated osteoclasts in the presence of CSF‐1 and RANKL, and impaired chemotaxis. Whereas the chemotaxis of CD44−/− cells to various chemoattractants is almost completely abrogated, a differential effect is seen with the OPN−/− cells. Thus, OPN−/− cells migrate normally towards CSF‐1 but not towards fMLP and MCP‐1, which signal through G‐protein coupled receptors (GPCRs). That the GPCR‐mediated migration is dependent upon the level of cell‐surface CD44 is indicated by the reduced cell‐surface expression of CD44 in OPN−/− cells and a comparable impairment in the chemotaxis of CD44+/− cells. Although chemotaxis of OPN−/− cells could be rescued by an OPN substratum, or by addition of high levels of OPN in solution, no response is evident with physiological levels of OPN, indicating a requirement for the CD44‐associated intracellular OPN in CD44 cell‐surface expression. These studies indicate, therefore, that the level of cell surface CD44 is critical for GPCR‐mediated chemotaxis by peritoneal macrophages and suggest that a novel intracellular form of OPN may modulate CD44 activities involved in these processes. J. Cell. Physiol. 198: 155–167, 2004.

Spiers Memorial Lecture. Lessons from biomineralization: comparing the growth strategies of mollusc shell prismatic and nacreous layers in Atrina rigida.

The mollusc shell prismatic layer of Atrina rigida is composed of an assemblage of large and relatively perfect single calcite crystals, embedded in an organic matrix. A key to elucidating basic mechanisms of mineralization is understanding the structures of the matrix, the mineral and the relations between them. The matrix that envelopes each prism (the inter-prismatic matrix) is composed mainly of glycine-rich proteins, while the matrix inside each prism (intra-crystalline matrix) is composed of a network of chitin fibers. Prisms grow by deposition of mineral particles on the chitin fibers. The mineral particles are associated with highly acidic proteins from the Asprich family, which presumably stabilize an amorphous mineral precursor. We infer that once in contact with the already formed crystalline prism, the particles crystallize by epitaxial nucleation. In nacre, sheets of beta-chitin are interspaced by silk-like proteins in a hydrated gel-like state. beta-Chitin forms a scaffold onto which the acidic proteins are adsorbed. Some of these are organized into a crystal nucleation site, where nucleation of aragonite, supposedly from colloidal amorphous calcium carbonate particles, is induced. Comparing the mechanisms of growth of the nacreous and prismatic layers can help to understand the underlying strategies of formation of mineralized structures.

Colocalization of intracellular osteopontin with CD44 is associated with migration, cell fusion, and resorption in osteoclasts.

Although osteopontin (OPN) is recognized generally as a secreted protein, an intracellular form of osteopontin (iOPN), associated with the CD44 complex, has been identified in migrating fibroblastic cells. Because both OPN and CD44 are expressed at high levels in osteoclasts, we have used double immunofluorescence analysis and confocal microscopy to determine whether colocalization of these proteins has functional significance in the formation and activity of osteoclasts. Analysis of rat bone marrow‐derived osteoclasts revealed strong surface staining for CD44 and β1‐ and β3‐integrins, whereas little or no staining for OPN or bone sialoprotein (BSP) was observed in nonpermeabilized cells. In permeabilized perfusion osteoclasts and multinucleated osteoclasts, staining for OPN and CD44 was prominent in cell processes, including filopodia and pseudopodia. Confocal microscopy revealed a high degree of colocalization of OPN with CD44 in motile osteoclasts. In cells treated with cycloheximide (CHX), perinuclear staining for OPN and BSP was lost, but iOPN staining was retained within cell processes. In osteoclasts generated from the OPN‐null and CD44‐null mice, cell spreading and protrusion of pseudopodia were reduced and cell fusion was impaired. Moreover, osteoclast motility and resorptive activity were significantly compromised. Although the area resorbed by OPN‐null osteoclasts could be rescued partially by exogenous OPN, the resorption depth was not affected. These studies have identified an intracellular form of OPN, colocalizing with CD44 in cell processes, that appears to function in the formation and activity of osteoclasts.

Binding of bone sialoprotein, osteopontin and synthetic polypeptides to hydroxyapatite.

The phosphorylated acidic glycoproteins bone sialoprotein (BSP) and osteopontin (OPN) bind to hydroxyapatite (HA) crystals and may be involved in the regulation of bone mineralization. The HA-binding properties of these proteins have been attributed to glutamic acid-rich sequences in BSP and aspartic acid-rich sequences in OPN. The present study examines the roles of these polycarboxylate sequences in the binding of BSP and OPN to HA. Porcine BSP, OPN and the synthetic polypeptides poly-L-glutamic acid [Poly(Glu)] and poly-L-aspartic acid [Poly(Asp)] were labeled with fluorescein isothiocyanate and their binding to HA determined by fluorimetry. From the binding isotherms, dissociation constants (KDs) for all the reagents tested were determined to be in the micromolar range. The saturation binding capacities of HA for Poly(Glu), Poly(Asp), BSP and OPN were similar (500–600 μg/m2). To investigate the role of glutamic acid-rich and aspartic acid-rich sequences in the binding to HA of BSP and OPN, respectively, competitive binding studies with Poly(Glu) and Poly(Asp) were performed. Poly(Glu) was able to displace a maximum of 100% of Poly(GIu), 81% of OPN, 68% of BSP and 65% of Poly(Asp). Poly(Asp) was able to displace a maximum of 100% of Poly(Glu), 99% of Poly(Asp), 95% of OPN and 89% of BSP. These results are consistent with the view that BSP and OPN bind to HA via their polycarboxylate sequences, but suggest a complex mode of interaction between polyelectrolytes and ionic crystals.

Bone sialoprotein-collagen interaction promotes hydroxyapatite nucleation.

In bone, hydroxyapatite (HA) crystals are deposited onto the type I collagen scaffold by a mechanism that has yet to be elucidated. Bone sialoprotein (BSP) is an acidic phosphoprotein that is expressed at high levels in mineralized tissues, capable of binding type I collagen, and nucleating HA. Both bone-extracted and recombinant BSP (rBSP) bind with equal affinity to collagen. The nature of the BSP-collagen interaction and its role in HA nucleation are not known. We have used a solid-phase binding assay and affinity chromatography to characterize the BSP-collagen interaction. rBSP-binding affinities of triple-helical and fibrillar type I collagen were similar (K(D) approximately 13 nM), while that of heat-denatured type I collagen was lower (K(D) approximately 44 nM), indicating the importance of triple-helical structure in binding BSP. Pepsin treatment of collagen had no effect on rBSP binding, demonstrating that the telopeptides of collagen are not involved. The majority of collagen-bound rBSP was eluted by acetonitrile, indicating that hydrophobic interactions are principally responsible for binding. Using an HA-nucleation assay, it was shown that rBSP is ten-fold more potent in reconstituted fibrillar collagen gels than in agarose gels. Nucleating potency of a non-collagen-binding, HA-nucleating peptide [rBSP(134-206)] showed no difference in the two gel systems. The work here shows that optimal binding of rBSP requires collagen to be in a native, triple-helical structure, does not require the telopeptides, and is stabilized by hydrophobic interactions. Upon binding to collagen, rBSP displays an increase in nucleation potency, implying a co-operative effect of BSP and collagen in mineral formation.

Identification of the Type I Collagen-binding Domain of Bone Sialoprotein and Characterization of the Mechanism of Interaction

Bone sialoprotein (BSP) is an anionic phosphorylated glycoprotein that is expressed almost exclusively in mineralized tissues and has been shown to be a potent nucleator of hydroxyapatite formation. The binding of BSP to collagen is thought to be important for the initiation of bone mineralization and in the adhesion of bone cells to the mineralized matrix. Using a solid phase assay, we have investigated the interaction between BSP and collagen. Initial studies showed that raising the ionic strength, decreasing the pH below 7, or introducing divalent cations diminishes but does not abolish the binding of BSP to collagen, indicating that the interaction is only partly electrostatic in nature. Both bone-extracted and recombinant (r)BSP exhibited similar binding affinities, indicating that post-translational modifications are not critical for binding. To identify the collagen-binding domain, recombinant peptides of BSP were studied. Peptide rBSP-(1–100) binds to type I collagen with an affinity similar to that of full-length rBSP, whereas peptides containing the sequences 99–201 or 200–301 do not bind. Further studies showed that rBSP-(1–75) competitively inhibits the binding of rBSP-(1–100), whereas rBSP-(21–100) inhibits binding to a lesser extent, and rBSP-(43–100) does not inhibit binding. These results suggest that the collagen-binding site of rat BSP is within the sequence 21–42, with residues N-terminal of this region likely also involved. This site was confirmed by the demonstration of collagen-binding activity of a synthetic peptide corresponding to residues 19–46. The collagen-binding domain, which is highly conserved among species, is enriched in hydrophobic residues and lacks acidic residues. We conclude that residues 19–46 of BSP represent a novel collagen-binding site.

Functional Analysis of Bone Sialoprotein: Identification of the Hydroxyapatite-nucleating and Cell-binding Domains by Recombinant Peptide Expression and Site-directed Mutagenesis

Mammalian bone sialoprotein (BSP) is a mineralized tissue-specific protein containing an RGD (arginine-glycine-aspartic acid) cell-attachment sequence and two distinct glutamic acid (glu)-rich regions, with each containing one contiguous glu sequence. These regions have been proposed to contribute to the attachment of bone cells to the extracellular matrix and to the nucleation of hydroxyapatite (HA), respectively. To further delineate the domains responsible for these activities, porcine BSP cDNA was used to construct expression vectors coding for two partial-length recombinant BSP peptides: P2S (residues 42-87), containing the first glutamic acid-rich domain; and P1L (residues 69-300), containing the second glutamic acid-rich region and the RGD sequence. These peptides were expressed in Escherichia coli as his-tag fusion proteins and purified by nickel affinity columns and FPLC chromatography. Digestion with trypsin released the his-tag fusion peptide, which generated P2S-TY (residues 42-87) and P1L-TY (residues 132-239). Using a steady-state agarose gel system, P2S-TY promoted HA nucleation, whereas P2S, P1L, and P1L-TY did not. This implies that the minimum requirement for nucleation of HA resides within the amino acid sequence of the first glutamic acid-rich domain, whereas the second glutamic acid-rich domain may require posttranslational modifications for activity. P1L, but not P2S, promoted RGD-mediated attachment of human gingival fibroblasts in a manner similar to that of native BSP. Deletion of the RGD domain or conversion of it to RGE (arginine-glycine-glutamic acid) abolished the cell-attachment activity of P1L. This suggests that, at least for human gingival fibroblasts, the major cell-attachment activity in the recombinant BSP peptides studied (residues 42-87 and 69-300) requires the RGD sequence located at the C-terminal domain.

Osteopontin Posttranslational Modifications, Possibly Phosphorylation, Are Required for In Vitro Bone Resorption but Not Osteoclast Adhesion

Osteopontin (OPN), a phosphorylated bone matrix glycoprotein, is an Arg-Gly-Asp (RGD)-containing protein that interacts with integrins and promotes in vitro attachment of a number of cell types, including osteoclasts. Gene knockout experiments support the idea that OPN is important in osteoclastic activity. We hypothesize that posttranslational modifications (PTMs) of OPN can influence its physiological function. Previous studies have suggested that phosphorylation of OPN and bone sialoprotein (BSP) is necessary for promoting osteoclast adhesion. However, no reports have explored the importance of phosphoserines and other PTMs in OPN-promoted bone resorption. To study this question, we determined the activities of different forms of OPN and BSP in three in vitro assays: attachment of osteoclasts; formation of actin rings; and bone resorption. For each assay, cells were incubated for 4-24 h, in the presence or absence of RGDS or RGES peptides, to test the involvement of integrin binding. In addition to OPN, activities of milk OPN (fully phosphorylated) and recombinant OPN (rOPN, no phosphate) were compared. We purified two forms of OPN (OPN-2 and OPN-5), which differ in the level of phosphorylation, and compared their activities. For comparison, the activities of BSP and recombinant BSP (rBSP) were determined. All forms of OPN, including rOPN, significantly increased attachment of tartrate-resistant acid phosphatase (TRAP)-positive osteoclasts. BSP and rBSP also promoted cell attachment. After 4 h of incubation, the proportion of cells with actin rings was increased with OPN, milk OPN, and BSP. In the presence of RGDS peptide, osteoclast retraction and the disruption of actin rings were observed, whereas no effect was seen with RGES. In the resorption assay, the number of pits and the total resorbed area per slice were increased in the presence of OPN, milk OPN, and BSP. As in other assays, the OPN enhancement of resorption was inhibited by RGDS, but not RGES, peptides. Significantly, rOPN and rBSP did not promote bone resorption. OPN-5 promoted resorption to a greater extent than OPN-2, and milk OPN significantly stimulated resorption to a greater extent than OPN. Our data suggest that: (1) the RGD sequence of OPN is essential in OPN-mediated cell attachment, actin ring formation, and bone resorption; and (2) some form of PTM, possibly phosphorylation, is necessary for in vitro osteoclastic bone resorption, but not for cell attachment and actin ring formation.

The flexible polyelectrolyte hypothesis of protein−biomineral interaction

Biomineralization is characterized by a high degree of control over the location, nature, size, shape, and orientation of the crystals formed. For many years, it has been widely believed that the exquisitely precise nature of crystal formation in biological tissues is the result of stereochemically specific interactions between growing crystals and extracellular matrix proteins. That is, the ability of many mineralized tissue proteins to adsorb to particular faces of biominerals has been attributed to a steric and electrical complementarity between periodic regions of the polypeptide chain and arrays of ions on the crystal face. In recent years, however, evidence has accumulated that many mineral-associated proteins lack periodic structure even when adsorbed to crystals. It also appears that protein-crystal interactions involve a general electrostatic attraction rather than arrays of complementary charges. In the present work, we review these studies and present some relevant new findings involving the mineral-modulating phosphoprotein osteopontin. Using molecular dynamics simulations, we show that the adsorption of osteopontin peptides to hydroxyapatite crystals does not involve a unique conformation of the peptide molecule, and that the adsorbed peptides are not aligned with rows of Ca(2+) ions on the crystal face. Further, we show that the interface between osteopontin peptides and calcium oxalate monohydrate crystals consists of peptide regions of high electronegativity and crystal faces of high electropositivity. Collectively, the above-mentioned studies suggest that interactions between mineral-modulating proteins and biologically relevant crystals are primarily electrostatic in nature, and that molecular disorder assists these proteins in forming multiple bonds with cations of the crystal face.

Determination of the Hydroxyapatite-Nucleating Region of Bone Sialoprotein

Bone sialoprotein (BSP) was shown to be a potent nucleator of hydroxyapatite (HA) in a steady-state agarose gel system (Hunter and Goldberg, 1993, PNAS 90: 8562). Nucleation of HA was also demonstrated with the homopolymer poly-glutamic acid but not with poly-aspartic acid or osteopontin. Since BSP contains contiguous sequences of glutamic acid, it is reasonable to suggest that the HA-nucleating activity of BSP resides within these regions. Purified porcine BSP was treated with trypsin and digests fractionated by gel filtration. In addition to small peptides (P3-5), two peptides of 38 kDa (P1) and 25 kDA (P2) were recovered, and after characterization assigned to the regions within BSP encompassing residues 133-272 (P1) and 42-125 (P2). Each of these peptides contained one of the two glutamic acid-rich regions of porcine BSP. In the steady-state agarose gel system, BSP, P1 and P2 induced HA formation, whereas the pooled small BSP-derived peptides (P3-5) did not. Analysis by circular dichroism spectroscopy revealed that the homopolymer poly-L-glutamic acid assumes a helical structure, while poly-L-aspartic acid does not. These findings suggest that the nucleating activity does not require intact molecules, that the nucleation of HA and BSP appears to require glutamic acid-rich sequences in a helical conformation and that there are two domains in porcine BSP that are each capable of nucleating HA.