Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Jul 1.
Published in final edited form as: Arterioscler Thromb Vasc Biol. 2013 May 2;33(7):1655–1662. doi: 10.1161/ATVBAHA.113.301523

Molecular basis of anti-angiogenic thrombospondin-1 type 1 repeat domain interactions with CD36

Philip A Klenotic 1,†, Richard C Page 1,†, Wei Li 2, Joseph Amick 1, Saurav Misra 1,*, Roy L Silverstein 3,*
PMCID: PMC3737738  NIHMSID: NIHMS485430  PMID: 23640500

Abstract

Objective

Anti-angiogenic activity of thrombospondin-1 (TSP-1) and related proteins is mediated by interactions between thrombospondin type I repeat (TSR) domains and the CLESH (CD36, LIMP-2, Emp sequence homology) domain of the endothelial cell receptor CD36. We sought to characterize key molecular determinants of the interaction between TSP1 TSR domains and the CD36 CLESH domain.

Approach and Results

Recombinant TSP-1 TSR2 and TSR(2,3) constructs inhibited microvascular endothelial cell (MVEC) migration, MVEC tube formation and vessel sprouting in aortic ring assays. Interaction with CD36-CLESH decoy peptides negated these effects. Mutational analyses identified a cluster of residues that confer positive charge to the TSR2 surface and mediate interaction with CD36 CLESH. Anti-angiogenic activity was significantly reduced by charge-neutralizing mutations of the Arg-Trp ladder in TSR2, but not TSR3. Additionally, I438 and K464 of TSR2 were shown to be required for CD36 CLESH binding to TSR2. A complementary acidic cluster within CD36 CLESH is also required for anti-angiogenic activity.

Conclusions

TSP-1 interacts with CD36 CLESH through electrostatic interactions mediated by a positively charged TSR2 surface and multiple negatively charged CD36 CLESH residues. Two key residues serve as specificity determinants that identify other TSR domains that interact with CD36 CLESH.

Keywords: thrombospondin-1, CD36, angiogenesis

The matricellular protein thrombospondin-1 (TSP-1) was the first endogenous protein inhibitor of angiogenesis to be identified.1 TSP-1, first isolated from human platelets,2 inhibits endothelial cell proliferation3 and migration and promotes apoptosis in the context of pro-angiogenic signals.4, 5 At early stages of dermal wound healing TSP-1 is released from platelet alpha granules into thrombi and the extracellular matrix to delay vascular remodeling.6 Within the tumor microenvironment, downregulation of TSP-1 is critical for many tumor cell types to continue unregulated cellular growth, metastasis and seeding to secondary sites.7 Among the five members of the thrombospondin family, only TSP-1 and TSP-2 have been shown to have anti-angiogenic activity to date. The functions of TSP-3 and TSP-4 are currently unclear. TSP-5, also termed cartilage oligomeric matrix protein (COMP), regulates attachment and survival of chondrocytes and potentially of other musculoskeletal cells,8-11 but no angiogenic functions for this protein have been reported.

TSP-1 and TSP-2 each contain three highly homologous thrombospondin type I repeat (TSR) domains. The other members of the thrombospondin family do not contain TSRs, emphasizing that TSR domains are important for anti-angiogenic function. Furthermore, several other anti-angiogenic proteins such as brain angiogenesis inhibitor-1 (BAI1) also contain TSR domains. The best-studied TSRs, the second and third TSR domains of TSP-1 (termed TSR(2,3)), bind and activate TGFβ,12-16 bind heparan sulfate proteoglycans and fibronectin,7 and promote endothelial cell apoptosis through the cell surface receptor CD36.17, 18 Recombinant TSR domains and TSR-derived synthetic peptides exhibit potent anti-angiogenic activity within tumors19 and attenuate tumor growth in mice.20-22 While TSR-derived peptides have been tested in phase II clinical trials, results are mixed.23 A more detailed understanding of the structural mechanisms utilized by TSR domains to regulate angiogenesis may inform the design of future therapies against pathological angiogenesis.

Two available crystal structures of TSP-1-TSR(2,3)11, 24 showcase a positively charged surface ridge within each TSR. These ridges have been proposed to serve as a recognition surface for TSP-1 and TSP-2 binding partners,25 most notably CD36. A ∼30-residue-long region within CD36 termed the CLESH (CD36, LIMP-2, Emp sequence homology) domain is sufficient for TSR binding.26 We and others have suggested that the negatively charged CLESH domain interacts with TSR domains via the positively charged surface ridge.11, 27 Moreover, we reported that phosphorylation of a CD36 threonine adjacent to the CLESH domain negatively regulates thrombospondin-1 binding.27 The basis and extent of TSR/CD36 CLESH interactions, however, remain poorly defined.

In the current study, we characterize the molecular basis of interaction between the TSR domains of TSP-1 and the CD36 CLESH domain. We show that only the second TSR domain (TSR2) of TSP-1 binds to the CD36 CLESH domain to inhibit MVEC angiogenic functions. We propose and validate key specificity determinants that mediate that interaction of TSR2 with CD36 CLESH and allow for the identification of other TSR domains that interact with CD36 CLESH. This study provides a framework to better understand the CD36 CLESH/TSR binding interface and the molecular determinants that mediate TSP-1 binding to CD36 at the endothelial cell surface.

Materials and Methods

Materials and Methods are available in the online-only Supplement.

Results

TSR2 Charge Neutralizing Mutations Abolish the Thrombospondin-1/CD36 CLESH Interaction

The second and third type I repeat domains of thrombospondin-1 (TSR2 and TSR3) were previously identified as the domains responsible for anti-angiogenic inhibition of MVEC-d migration.24, 28 Sequence analyses have also identified TSR domains within the anti-angiogenic proteins thrombospondin-2 (TSP-2), brain angiogenesis inhibitor 1 (BAI1) and a disintegrin and metalloproteinase with thrombospondin motifs 1 and 4 (ADAMTS-1 and ADAMTS-4). Two conserved arginine residues (Figure 1A) participate in cation-π stacking with neighboring tryptophan residues to form the CWR-layered core of each TSR domain.24, 25 These arginine residues confer a positive charge to one surface ridge of the CWR-layered core (Figure 1B). Based on electrostatic surface calculations and molecular modeling, we hypothesized that novel Arg-to-Met mutations could reduce the positive charge of the CWR-layered core surface ridge (Figure 1B; Supplemental Figure I) without compromising the structural integrity of the TSRs.

Figure 1.

Figure 1

Open in a new tab

Conserved arginine residues in TSR2 are required for the anti-migratory activity of TSR(2,3). A, Sequence alignment of the second and third TSR domains of TSR(2,3) against TSR-containing proteins with anti-angiogenic activity. Conserved arginine (cyan) and tryptophan (magenta) residues within the CWR-layered core are highlighted. B, APBS-calculated electrostatic surface representations of wild-type and Arg-to-Met mutant TSR2 or TSR3. Molecular surfaces are colored according to the solvent accessible surface potential from +5 kT (blue) to −5 kT (red). C, Modified Boyden chamber migration assay of MVEC-d detected by DAPI staining. MVEC-d were allowed to migrate toward low growth factor media (“Low GF”) or media containing a full complement of growth factors (“Normal GF”). MVEC-d migration assays performed in the presence of recombinant wild-type or Arg-to-Met mutant TSR(2,3) were allowed to migrate towards media containing a full complement of growth factors. Data represent mean±SD of three independent experiments. **P<0.001.

We analyzed the effect of individual Arg-to-Met mutations of TSR(2,3) on the inhibition of MVEC-d migration in a modified Boyden chamber assay. Few cells migrated when incubated against a low growth factor media chemoattractant containing 10% of normal growth factors (Figure 1C). Incubation against a normal growth factor media chemoattractant containing 100% of normal growth factors resulted in a five-fold increase in migration. Addition of recombinant TSR(2,3) to normal growth factor media reduced MVEC-d migration to a level similar to that observed with low growth factor media. Arg-to-Met mutations in the TSR2 domain (R440M and R442M) fully abrogated the inhibitory activity of TSR(2,3). In contrast, the equivalent Arg-to-Met mutations in the TSR3 domain (R497M or R499M) had no significant effect on the inhibitory activity of TSR(2,3).

Since the effect of Arg-to-Met mutations could be ascribed to disruptions in the structure of TSR2, we acquired solution NMR HSQC spectra for uniformly 15N-labeled wild-type, R440M- and R442M-TSR2. HSQC spectra for R440M- and R442M-TSR2 are well dispersed and very similar to the HSQC spectrum of wild-type TSR2 (Supplemental Figure II), verifying that the R440M and R442M mutations do not substantially alter the tertiary structure of TSR2.

TSR2 is the Critical TSR Required for CD36 CLESH Binding

To confirm that TSR2 is the minimal domain responsible for the CD36 CLESH-dependent anti-migratory activity of TSP-1, we conducted MVEC-d migration assays with TSR(2,3), TSR2 and TSR3 domain constructs. TSR2 attenuated MVEC-d migration to the same extent as the TSR(2,3) construct (Figure 2A). Attenuation of MVEC-d migration by TSR2 was reversed upon addition of a competing GST-CD36 CLESH fusion protein, whereas GST alone had no effect (Figure 2A). In comparison, TSR3 only mildly attenuated MVEC-d migration and this attenuation was insensitive to CD36 CLESH (Figure 2A).

Figure 2.

Figure 2

Open in a new tab

The CLESH domain of CD36 binds and reverses the anti-migratory activity of TSR2, but not TSR3. A, MVEC-d migration in a modified Boyden chamber toward low growth factor media (“Low GF”) or media containing a full complement of growth factors (“Normal GF”) compared to MVEC-d migration towards media containing a full complement of growth factors in the presence of recombinant TSR(2,3), TSR2, TSR2 and GST-CD36 CLESH (“TSR2 + CLESH”), TSR2 and GST (“TSR2 + GST”), TSR3 or TSR3 and GST-CD36 CLESH (“TSR3 + CLESH”). Migrated cells were detected by DAPI staining. B, CLESH-Ub or Ub alone were incubated in microtiter plate wells that were blocked (“Blocked Well”) or pre-treated with TSR(2,3), TSR2 or TSR3. After incubation and extensive washing, bound CD36 CLESH-Ub or Ub alone were detected by sequential incubations with anti-ubiquitin primary antibody, horseradish peroxidase-conjugated secondary antibody and TMB-ELISA substrate. Data represent mean±SD of three independent experiments. **P<0.001. N.S. denotes no significant difference.

We next carried out solid phase binding assays to directly measure binding of CD36 CLESH to TSR(2,3), TSR2 or TSR3. In a binding assay using CD36 CLESH-ubiquitin fusion protein (CLESH-Ub) probed with an anti-ubiquitin antibody, the CLESH-Ub fusion bound similarly to TSR2 and TSR(2,3) (Figure 2B). We observed minimal binding of CLESH-Ub to TSR3, or in a control experiment measuring binding of ubiquitin to TSR2.

TSR2, but not TSR3, Inhibits Angiogenesis in a CD36 CLESH Dependent Manner

Solid-phase binding assays and MVEC-d migration assays suggested that TSR2 exhibits CD36 CLESH-dependent anti-angiogenic activity while TSR3 does not. To evaluate this hypothesis we performed ex vivo angiogenesis assays that examined branching in MVEC-b tube formation29, 30 and assessed vessel sprouting of aortic rings.31 MVEC-b tube formation assays with TSR2 reduced tube formation to levels similar to that observed in low growth factor-containing media (Figure 3A and Supplementary Figure III). Tube formation was fully restored by the addition of CD36 CLESH (Figure 3A). In comparison, TSR3 slightly inhibited MVEC-b tube formation; however, this inhibition was insensitive to CD36 CLESH (Figure 3A). TSR2 also significantly impaired vessel sprouting in an aortic ring assay (Figure 3B and Supplementary Figure IV). This inhibitory activity was reversed upon addition of CD36 CLESH (Figure 3B). TSR3 had no significant effect upon aortic ring vessel sprouting and exhibited no CD36 CLESH dependence (Figure 3B).

Figure 3.

Figure 3

Open in a new tab

TSR2, but not TSR3, exhibits CD36-CLESH dependent anti-angiogenic activity in MVEC tube formation and aortic ring assays. A, MVECs-b were seeded onto Matrigel and incubated for 6 hours with a low growth factor media (“Low GF”), media with full complement of growth factors (“Normal GF”), or full growth factor media in the presence of TSR2, TSR2 and CD36 CLESH-Ub (“TSR2 + CLESH”), TSR3 or TSR3 and CD36 CLESH-Ub (“TSR3 + CLESH”). Tube formation was assessed by counting the number of branch points per 10× field as detected by phase contrast microscopy. B, Aortic rings were cultured in Matrigel with media containing a full complement of growth factors (“Normal GF”), or full media containing recombinant wild-type TSR2, TSR2 and CD36 CLESH-Ub (“TSR2 + CLESH”), TSR3 or TSR3 and CD36 CLESH-Ub (“TSR3 + CLESH”). Total vessel formation per aortic ring after 6 days was assessed by phase contrast microscopy. Data represent mean±SD of three independent experiments. **P<0.001. *P<0.01. N.S. denotes no significant difference.

Conversion of TSR2 to TSR3 Eliminates CD36 CLESH Effects

We carried out solution NMR HSQC titrations to identify TSR2 residues that interact with CD36 CLESH or are located near the CLESH-binding site (Figure 4A). Such residues are predicted to exhibit chemical shift perturbations (CSPs) upon addition of CLESH-Ub. As a control, we observed no significant CSPs induced by ubiquitin alone (Figure 4A). Residues that undergo significant CSPs (greater than 3σ above noise) cluster along the surface ridge of the CWR-layered core and a nearby pocket (Figure 4B). The NMR data, in combination with migration assay results for wild-type and Arg-to-Met mutant TSR(2,3) constructs, confirm that the positively charged surface ridge of the TSR2 CWR-layered core is a key determinant for CD36 CLESH binding.

Figure 4.

Figure 4

Open in a new tab

TSR3 lacks key determinants required for CD36 CLESH binding. A, Chemical shift perturbation (CSP, Δδ) histogram for TSR2 titrated with a 10× molar excess of CLESH-Ub (green) or Ub (black). Dashed lines indicate the cutoff values for residues exhibiting CSPs greater than 3σ or 6σ above the noise. B, Significant CSPs mapped to the structure of TSR2 (white) shown in cartoon representation. Residues are colored corresponding to CSPs greater than 3σ (cyan) or 6σ (blue) above the noise. C, Sequence alignment of the second and third TSR domains of TSR(2,3). Residues I438 and K464, which confer a positive charge to TSR2 and corresponding residues in TSR3 are highlighted (black). D, APBS-calculated electrostatic surface representations of wild-type TSR2, I438Q/K464Q mutant TSR2, wild-type TSR3 and Q495I/Q521K mutant TSR3 colored according to the solvent accessible surface potential from +5 kT (blue) to −5 kT (red). E, MVEC-ds were allowed to migrate toward low growth factor media (“Low GF”), media containing a full complement of growth factors (“Normal GF”), or full media containing recombinant wild-type TSR(2,3), I438Q/K464Q-TSR(2,3), Q495I/Q521K-TSR3, or BAI1-TSR3 with or without CD36 CLESH. F, MVEC-bs plated onto Matrigel were incubated for 6 hours against low growth factor media (“Low GF”), media containing a full complement of growth factors (“Normal GF”), or full growth factor media in the presence of Q495I/Q521K-TSR3 (“TSR3 Q495I/Q521K”) or Q495I/Q521K-TSR3 and CD36 CLESH (“TSR3 Q495I/Q521K + CLESH”). Tube formation was assessed by counting the number of branch points er 10× field as detected by phase contrast microscopy. G, Aortic rings were cultured in Matrigel with full growth factor media (“Normal GF”), or full growth factor media with TSR2, Q495I/Q521K-TSR3 (“TSR3 Q495I/Q521K”) or Q495I/Q521K-TSR3 and CD36 CLESH (“TSR3 Q495I/Q521K + CLESH”). Total vessel formation per aortic ring after 6 days was assessed by phase contrast microscopy. Data represent mean±SD of three independent experiments. **P<0.001. N.S. denotes no significant difference.

We compared putative CLESH-interacting TSR2 residues that were identified in our NMR HSQC experiments with the sequence of TSR3. We identified two non-conserved residues, I438 and K464 (Figure 4C). These two amino acids contribute to the overall positive character of the CLESH binding surface of TSR2 (Figure 4D). The corresponding residues in TSR3, Q495 and Q521, result in a more neutral surface (Figure 4D). We therefore introduced dual I438Q/K464Q mutations into TSR(2,3) to neutralize the CLESH-binding surface of TSR2 (Figure 4D). We also introduced dual Q495I/Q521K mutations into TSR3 to increase the overall positive character of TSR3 and produce a surface with similar charge characteristics as TSR2 (Figure 4D).

While wild-type TSR(2,3) effectively inhibited MVEC-d migration, TSR(2,3)-I438Q/K464Q was only partly effective (Figure 4E). Furthermore, while CD36 CLESH blocked the inhibition of MVEC-d migration by wild-type TSR(2,3), it had no effect on TSR(2,3)-I438Q/K464Q (Figure 4E). In contrast, TSR3-Q495I/Q521K effectively inhibited migration and this inhibitory activity was sensitive to CD36 CLESH (Figure 4E). These results suggest that Q495I/Q521K mutations within TSR3 were sufficient to confer CD36 CLESH sensitivity to a TSR domain. We tested whether this “converted” TSR domain also exhibits anti-angiogenic activity similar to wild-type TSR2 (Figure 3). MVEC-b tube formation assays (Figure 4F) and aortic ring assays (Figure 4G) confirmed the CD36 CLESH dependent anti-angiogenic activity of TSR3-Q495I/Q521K. These results suggest that the I438/K464 and Q495/Q521 pockets contribute significantly to the surface charge characteristics of the TSR domains, and that these residues represent key molecular determinants for CD36 CLESH-dependent anti-angiogenic activity.

Sequence alignments of TSR domains from proteins with anti-angiogenic activity (Figure 1A and Supplemental Figure V) identify TSP-2 TSR2 and BAI1 TSR3 as positively charged TSR domains similar to TSP-1 TSR2. We tested whether BAI1 TSR3 exhibits anti-migratory activity in a CD36 CLESH-dependent manner (Figure 4E). BAI1 TSR3 effectively inhibited MVEC-d migration in a manner similar to TSP-1 TSR(2,3). Additionally, CD36 CLESH significantly reduced BAI TSR3 inhibition of MVEC-d migration.

Acidic Residues in CD36 CLESH are Critical for Interaction with TSR2

The CD36 CLESH domain has four acidic residues, E101, D106, E108 and D109 (Figure 5A) that potentially interact with the positively charged CLESH-binding surface of TSR2. We therefore tested individual and multiple alanine mutations of these amino acids in MVEC-d migration assays. While wild-type CD36 CLESH blocked the inhibition of MVEC-d migration by TSR2 (Figure 5A “TSR2 + CLESH”), the E101A-, D106A-, E108A- and D109A-CD36 CLESH mutants only partly rescued MVEC-d migration (Figure 5B). Furthermore, the triple mutant D106A/E108A/D109A- and quadruple mutant E101A/D106A/E108A/D109A-CD36 CLESH did not rescue MVEC migration at all (Figure 5B). In combination with NMR binding studies (Figure 4), these data show that the CD36 CLESH acidic side-chains mediate the interaction between CD36 CLESH and the positively charged surface of TSR2.

Figure 5.

Figure 5

Open in a new tab

Negatively charged CD36 CLESH carboxylate side-chains are critical for the interaction with TSR2. A, Amino acid sequence of the extended CLESH domain (residues 81-117) including the T92 phosphorylation site (underlined) and negatively charged carboxylate-containing residues (highlighted, numbered). The canonical CLESH domain (residues 93-117) is boxed. B, MVEC-d migration assay detected by DAPI staining. MVEC-ds migrated toward low growth factor media (“Low GF”), media containing a full complement of growth factors (“Normal GF“), full media containing recombinant wild-type TSR2, or full media containing recombinant wild-type TSR2 and a mutant GST-CD36 CLESH construct (mutations shown in parentheses). Normalized MVEC-d migration values are expressed relative to the average migration observed towards full media (1.00 = 111 cells per 10× field). Data represent mean±SD of three independent experiments. *P<0.01.

Position 92 of the Extended CD36 CLESH Domain Regulates Interaction with TSR2

Our lab previously demonstrated that phosphorylation of the extended CD36 CLESH domain at T92 (Figure 5A) significantly blocked TSP-1 binding in an in vitro ELISA assay.27 Mutagenesis of CD36 suggested a steric component to the inhibition of CD36/TSP-1 interactions by T92 phosphorylation.27 To further explore this possibility we performed MVEC-d migration assays in the presence of TSR2 and CD36 CLESH in which T92 is substituted by tryptophan, glutamate or arginine, amino acids with bulky side-chains and a range of charge states. In comparison to wild-type CD36 CLESH, T92W-, T92E- or T92R-CD36 CLESH only partly restored MVEC-d migration blocked by TSR2 (Figure 5B). The similar effect of neutral, negatively charged and positively charged bulky side-chain substitutions at T92 suggests that phosphorylation of T92 may sterically rather than electrostatically interfere with CD36 CLESH/TSP-1 interactions.

Molecular Model of the TSR2/CD36 CLESH Complex

Our MVEC-d migration assays and in vitro ELISA and NMR binding studies provide residue-specific restraints that can be used to model the interaction between TSR2 and CD36 CLESH. We prepared a list of ambiguous restraints in which each TSR2 residue found by migration assays, ELISA or NMR to contribute to CD36 CLESH binding was restrained to a distance of 3.5Å or less from E101, D106, E108 or D109 of CD36 CLESH. Complementary restraints required these CD36 CLESH residues to be positioned 3.5Å or less from TSR2 residues found to contribute to CD36 CLESH binding. We utilized these ambiguous restraints as an input for a rigid body/torsion angle dynamics simulated annealing protocol to generate a model of the TSR2/CD36 CLESH complex. Consistent with the results of MVEC migration assays for the R440M, R442M, I438Q and K464Q TSR(2,3) mutants, the model suggests that the interaction between CD36 and TSP-1 is mediated by extensive electrostatic contacts between negatively charged CD36 CLESH carboxylate side-chains and residues that confer a positive charge to TSR2 (Figure 6).

Figure 6.

Figure 6

Open in a new tab

Molecular model of the TSR2/CD36 CLESH complex. A, Cartoon representation of the model TSR2/CD36 CLESH complex. B, Lowest energy model of the TSR2/CD36 CLESH complex with TSR2 shown as a surface representation. C, Ensemble of the ten lowest energy models for the TSR2/CD36 CLESH complex with TSR2 shown as a surface representation. Modified CPK coloring is used throughout A, B and C for nitrogen (blue), oxygen (red), sulfur (yellow) and carbon atoms of CD36 CLESH (pink) and TSR2 (white). Carbon atoms of TS R2 residues that participate in CLESH binding are colored according to identification by NMR (cyan), mutagenesis (green) or both (orange).

Discussion

While the interactions between TSP-1 TSR domains and the CD36 CLESH domain are central to the CD36-dependent anti-angiogenic and anti-migratory activities of TSP-1,26, 27 the molecular basis for the TSP-1/CD36 CLESH interaction has been poorly defined to date. In this study we combined structural and sequence data, mutagenesis, protein-protein interaction experiments and in vitro functional assays to characterize key determinants that regulate TSR interactions with CD36 CLESH.

Our NMR experiments identified a cluster of TSR2 surface residues that were perturbed upon addition of CD36 CLESH, suggesting a direct contact with the CD36 CLESH domain. Corneal pocket and endothelial cell migration assays previously implicated these residues in TSP-1 anti-angiogenic activity.17, 32 Using molecular modeling we identified an allowed side-chain rotamer for methionine that places the methionine δ-sulfur atom in a position for S-π interactions33-35 and preserves the π-stacking arrangement with neighboring tryptophan indole rings. The arginine to methionine substitutions within the CWR-layered cores of TSR2 and TSR3 likely neutralize the charged WR stacking motif surface ridges. We indeed observed that these mutations disrupt TSR2 interactions with CD36 CLESH.

In contrast to the highly conserved WR-stacking motif, an “I-K” positively charged patch on TSR2 formed by I438 and K464 serves as a differentiating factor between TSR2 and TSR3. The “I-K” patch is a key determinant that can be used to identify CD36 CLESH-interacting TSR domains on the basis of amino acid sequence (Supplemental Figure V). We hypothesize that among the five TSR-containing proteins with known anti-angiogenic activity, only TSP-1 TSR2, TSP-2 TSR2 and BAI1 TSR3 contain the necessary determinants for interacting with CD36 CLESH. These three TSR domains contain residues with side-chains that would confer a positive charge to the “I-K” patch, while other TSR domains within these proteins contain neutral patches that are likely unsuitable for CD36 CLESH binding. The TSR-containing fragment of BAI1 was previously identified as an inhibitor of intracranial glioma growth and progression through a CD36-dependent mechanism.36 Our “I-K” patch sequence-based categorizations predict that BAI1 TSR3 is responsible for BAI1 anti-angiogenic activity, consistent with previous reports from our lab and others.37 Indeed, we found that BAI1 TSR3 suppressed endothelial cell migration, and this inhibition was partly reversed by CD36 CLESH (Figure 4E). In contrast, the TSR domains of ADAMTS-1 and ADAMTS-4 contain sequence differences that neutralize both the “I-K” patch and the WR-stacking motif, and most likely do not interact with CD36 CLESH.

Our modeling suggests that the CD36 CLESH sequence binds in an extended conformation along the positively charged TSR surface. Our NMR and mutagenesis data show that both electrostatic and steric determinants (e.g. the phosphorylation state of T92) regulate TSR/CLESH binding. Interestingly, NMR titrations of 15N-CD36 CLESH with TSR2 suggest that CD36 CLESH remains unstructured and dynamic in the presence of TSR2 (data not shown). We speculate that CD36 CLESH may interact with TSR2 as an ensemble of conformations that satisfy overall electrostatic complementarity without enforcing rigid pairwise interactions between specific CD36 CLESH and TSR2 residues. The distributed electrostatic nature of the TSR2/CD36 CLESH interaction is consistent with the relatively small chemical shift perturbations observed in our NMR experiments. Such a “fuzzy interaction” hypothesis38 requires further investigation. However, it is noteworthy that the circulating protein HRGP also contains a CLESH-like domain that interacts with TSP-1 TSR domains. While the HRGP CLESH contains a similar number of acidic residues as the CD36 CLESH, its acidic residues are located in a different arrangement than in CD36.39

Interestingly, TSP1-TSR3 exhibited mild but reproducible anti-angiogenic activity that is independent of CD36 CLESH. The mechanism by which TSR3 exerts this effect is not known. TSR3 may interact with a distinct region of CD36, with a distinct endothelial receptor or with an extracellular signaling factor. The characterization of the mode of action of TSR3 will require future study.

The identification of TSR2 as the source of CD36 CLESH-mediated anti-angiogenic activity is relevant for the development of thrombospondin-based therapeutics. A recombinant TSP-1 construct containing the TSR1, TSR2 and TSR3 domains19, 40 inhibited growth of human pancreatic cancer cells in an orthotopic mouse model.40 However, this construct contains multiple epitopes that interact with distinct signaling pathways including activation of TGF-β via the RFK motif located between TSR1 and TSR2.12-16 Peptidomimetics modeled on sequences within TSP-1 TSR2 had limited success due to rapid clearance.22 While these peptide therapeutics were based on putative CD36-interacting epitopes, they most likely act through an entirely different pathway by activating TGF-β.22 We hypothesize that these differences are due to the unusual structure of the TSR domains and the inability of small peptides to fully recapitulate the positively charged surface ridge required for binding to CD36 CLESH.

The results of the present study identify molecular determinants that regulate the CD36 CLESH-dependent anti-angiogenic acitivty of TSP-1. The identification and characterization of the anti-angiogenic properties of recombinant, bacterially expressed TSR domains like TSP1-TSR2 and BAI1-TSR3 may contribute further to the development of TSR-based therapies in malignant gliomas and other cancers.

Supplementary Material

3
4

Significance Statement.

By interacting with the cell surface receptor CD36, the matricellular protein thrombospondin-1 (TSP-1) exhibits potent anti-angiogenic activity. Downregulation of TSP-1 within tumor microenvironments promotes unregulated cellular growth and metastasis. Previous studies identified an interaction between the TSP-1 type I repeat (TSR) domains and CD36. Here we identify the specific molecular determinants within the TSP-1 TSR domains and the CLESH domain of CD36 that mediate the TSP-1/CD36 interaction. Our results provide a framework for identifying which TSR domains within other proteins, such as thrombospondin-2 (TSP-2) and brain angiogenesis inhibitor 1 (BAI1), may also regulate angiogenesis by interacting with CD36. These findings give molecular insights into the anti-angiogenic effects of TSP-1 and present a basis for engineering anti-angiogenic or antitumor therapeutics based on TSR domains.

Acknowledgments

Sources of Funding: The authors acknowledge financial support from the US National Institutes of Health (grant R01-GM080271 to SM and grant R01-HL67839 to RLS). RCP was supported by a National Institutes of Health Postdoctoral Training Grant in Vascular Biology (T32-HL007914).

Footnotes

Disclosures: The authors declare no conflicts of interest.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Good DJ, Polverini PJ, Rastinejad F, Le Beau MM, Lemons RS, Frazier WA, Bouck NP. A tumor suppressor-dependent inhibitor of angiogenesis is immunologically and functionally indistinguishable from a fragment of thrombospondin. Proc Natl Acad Sci U S A. 1990;87:6624–6628. doi: 10.1073/pnas.87.17.6624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Lawler JW, Slayter HS, Coligan JE. Isolation and characterization of a high molecular weight glycoprotein from human blood platelets. J Biol Chem. 1978;253:8609–8616. [PubMed] [Google Scholar]
  • 3.Bagavandoss P, Wilks JW. Specific inhibition of endothelial cell proliferation by thrombospondin. Biochem Biophys Res Commun. 1990;170:867–872. doi: 10.1016/0006-291x(90)92171-u. [DOI] [PubMed] [Google Scholar]
  • 4.Short SM, Derrien A, Narsimhan RP, Lawler J, Ingber DE, Zetter BR. Inhibition of endothelial cell migration by thrombospondin-1 type-1 repeats is mediated by beta1 integrins. J Cell Biol. 2005;168:643–653. doi: 10.1083/jcb.200407060. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Isenberg JS, Martin-Manso G, Maxhimer JB, Roberts DD. Regulation of nitric oxide signalling by thrombospondin 1: implications for anti-angiogenic therapies. Nat Rev Cancer. 2009;9:182–194. doi: 10.1038/nrc2561. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Kyriakides TR, Maclauchlan S. The role of thrombospondins in wound healing, ischemia, and the foreign body reaction. J Cell Commun Signal. 2009;3:215–225. doi: 10.1007/s12079-009-0077-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Ren B, Yee KO, Lawler J, Khosravi-Far R. Regulation of tumor angiogenesis by thrombospondin-1. Biochim Biophys Acta. 2006;1765:178–188. doi: 10.1016/j.bbcan.2005.11.002. [DOI] [PubMed] [Google Scholar]
  • 8.Chen FH, Thomas AO, Hecht JT, Goldring MB, Lawler J. Cartilage oligomeric matrix protein/thrombospondin 5 supports chondrocyte attachment through interaction with integrins. J Biol Chem. 2005;280:32655–32661. doi: 10.1074/jbc.M504778200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Riessen R, Fenchel M, Chen H, Axel DI, Karsch KR, Lawler J. Cartilage oligomeric matrix protein (thrombospondin-5) is expressed by human vascular smooth muscle cells. Arterioscler Thromb Vasc Biol. 2001;21:47–54. doi: 10.1161/01.atv.21.1.47. [DOI] [PubMed] [Google Scholar]
  • 10.Briggs MD, Chapman KL. Pseudoachondroplasia and multiple epiphyseal dysplasia: mutation review, molecular interactions, and genotype to phenotype correlations. Hum Mutat. 2002;19:465–478. doi: 10.1002/humu.10066. [DOI] [PubMed] [Google Scholar]
  • 11.Tan K, Lawler J. The interaction of Thrombospondins with extracellular matrix proteins. J Cell Commun Signal. 2009;3:177–187. doi: 10.1007/s12079-009-0074-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Murphy-Ullrich JE, Poczatek M. Activation of latent TGF-beta by thrombospondin-1: mechanisms and physiology. Cytokine Growth Factor Rev. 2000;11:59–69. doi: 10.1016/s1359-6101(99)00029-5. [DOI] [PubMed] [Google Scholar]
  • 13.Schultz-Cherry S, Chen H, Mosher DF, Misenheimer TM, Krutzsch HC, Roberts DD, Murphy-Ullrich JE. Regulation of transforming growth factor-beta activation by discrete sequences of thrombospondin 1. J Biol Chem. 1995;270:7304–7310. doi: 10.1074/jbc.270.13.7304. [DOI] [PubMed] [Google Scholar]
  • 14.Schultz-Cherry S, Lawler J, Murphy-Ullrich JE. The type 1 repeats of thrombospondin 1 activate latent transforming growth factor-beta. J Biol Chem. 1994;269:26783–26788. [PubMed] [Google Scholar]
  • 15.Schultz-Cherry S, Ribeiro S, Gentry L, Murphy-Ullrich JE. Thrombospondin binds and activates the small and large forms of latent transforming growth factor-beta in a chemically defined system. J Biol Chem. 1994;269:26775–26782. [PubMed] [Google Scholar]
  • 16.Yee KO, Streit M, Hawighorst T, Detmar M, Lawler J. Expression of the type-1 repeats of thrombospondin-1 inhibits tumor growth through activation of transforming growth factor-beta. Am J Pathol. 2004;165:541–552. doi: 10.1016/s0002-9440(10)63319-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Dawson DW, Pearce SF, Zhong R, Silverstein RL, Frazier WA, Bouck NP. CD36 mediates the In vitro inhibitory effects of thrombospondin-1 on endothelial cells. J Cell Biol. 1997;138:707–717. doi: 10.1083/jcb.138.3.707. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Jimenez B, Volpert OV, Crawford SE, Febbraio M, Silverstein RL, Bouck N. Signals leading to apoptosis-dependent inhibition of neovascularization by thrombospondin-1. Nat Med. 2000;6:41–48. doi: 10.1038/71517. [DOI] [PubMed] [Google Scholar]
  • 19.Miao WM, Seng WL, Duquette M, Lawler P, Laus C, Lawler J. Thrombospondin-1 type 1 repeat recombinant proteins inhibit tumor growth through transforming growth factor-beta-dependent and -independent mechanisms. Cancer Res. 2001;61:7830–7839. [PubMed] [Google Scholar]
  • 20.Henkin J, Volpert OV. Therapies using anti-angiogenic peptide mimetics of thrombospondin-1. Expert Opin Ther Targets. 2011;15:1369–1386. doi: 10.1517/14728222.2011.640319. [DOI] [PubMed] [Google Scholar]
  • 21.Huang H, Campbell SC, Bedford DF, Nelius T, Veliceasa D, Shroff EH, Henkin J, Schneider A, Bouck N, Volpert OV. Peroxisome proliferator-activated receptor gamma ligands improve the antitumor efficacy of thrombospondin peptide ABT510. Mol Cancer Res. 2004;2:541–550. [PubMed] [Google Scholar]
  • 22.Recouvreux MV, Camilletti MA, Rifkin DB, Becu-Villalobos D, Diaz-Torga G. Thrombospondin-1 (TSP-1) analogs ABT-510 and ABT-898 inhibit prolactinoma growth and recover active pituitary transforming growth factor-beta1 (TGF-beta1) Endocrinology. 2012;153:3861–3871. doi: 10.1210/en.2012-1007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Markovic SN, Suman VJ, Rao RA, Ingle JN, Kaur JS, Erickson LA, Pitot HC, Croghan GA, McWilliams RR, Merchan J, Kottschade LA, Nevala WK, Uhl CB, Allred J, Creagan ET. A phase II study of ABT-510 (thrombospondin-1 analog) for the treatment of metastatic melanoma. Am J Clin Oncol. 2007;30:303–309. doi: 10.1097/01.coc.0000256104.80089.35. [DOI] [PubMed] [Google Scholar]
  • 24.Klenotic PA, Page RC, Misra S, Silverstein RL. Expression, purification and structural characterization of functionally replete thrombospondin-1 type 1 repeats in a bacterial expression system. Protein Expr Purif. 2011;80:253–259. doi: 10.1016/j.pep.2011.07.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Tan K, Duquette M, Liu JH, Dong Y, Zhang R, Joachimiak A, Lawler J, Wang JH. Crystal structure of the TSP-1 type 1 repeats: a novel layered fold and its biological implication. J Cell Biol. 2002;159:373–382. doi: 10.1083/jcb.200206062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Frieda S, Pearce A, Wu J, Silverstein RL. Recombinant GST/CD36 fusion proteins define a thrombospondin binding domain. Evidence for a single calcium-dependent binding site on CD36. J Biol Chem. 1995;270:2981–2986. doi: 10.1074/jbc.270.7.2981. [DOI] [PubMed] [Google Scholar]
  • 27.Chu LY, Silverstein RL. CD36 ectodomain phosphorylation blocks thrombospondin-1 binding: structure-function relationships and regulation by protein kinase C. Arterioscler Thromb Vasc Biol. 2012;32:760–767. doi: 10.1161/ATVBAHA.111.242511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Iruela-Arispe ML, Lombardo M, Krutzsch HC, Lawler J, Roberts DD. Inhibition of angiogenesis by thrombospondin-1 is mediated by 2 independent regions within the type 1 repeats. Circulation. 1999;100:1423–1431. doi: 10.1161/01.cir.100.13.1423. [DOI] [PubMed] [Google Scholar]
  • 29.Kubota Y, Kleinman HK, Martin GR, Lawley TJ. Role of laminin and basement membrane in the morphological differentiation of human endothelial cells into capillary-like structures. J Cell Biol. 1988;107:1589–1598. doi: 10.1083/jcb.107.4.1589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ponce ML. Tube formation: an in vitro matrigel angiogenesis assay. Methods Mol Biol. 2009;467:183–188. doi: 10.1007/978-1-59745-241-0_10. [DOI] [PubMed] [Google Scholar]
  • 31.Baker M, Robinson SD, Lechertier T, Barber PR, Tavora B, D'Amico G, Jones DT, Vojnovic B, Hodivala-Dilke K. Use of the mouse aortic ring assay to study angiogenesis. Nat Protoc. 2011;7:89–104. doi: 10.1038/nprot.2011.435. [DOI] [PubMed] [Google Scholar]
  • 32.Tolsma SS, Volpert OV, Good DJ, Frazier WA, Polverini PJ, Bouck N. Peptides derived from two separate domains of the matrix protein thrombospondin-1 have anti-angiogenic activity. J Cell Biol. 1993;122:497–511. doi: 10.1083/jcb.122.2.497. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Klingler TM, Brutlag DL. Discovering structural correlations in alpha-helices. Protein Sci. 1994;3:1847–1857. doi: 10.1002/pro.5560031024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Pal D, Chakrabarti P. Different types of interactions involving cysteine sulfhydryl group in proteins. J Biomol Struct Dyn. 1998;15:1059–1072. doi: 10.1080/07391102.1998.10509001. [DOI] [PubMed] [Google Scholar]
  • 35.Tatko CD, Waters ML. Investigation of the nature of the methionine-pi interaction in beta-hairpin peptide model systems. Protein Sci. 2004;13:2515–2522. doi: 10.1110/ps.04820104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Kaur B, Cork SM, Sandberg EM, Devi NS, Zhang Z, Klenotic PA, Febbraio M, Shim H, Mao H, Tucker-Burden C, Silverstein RL, Brat DJ, Olson JJ, Van Meir EG. Vasculostatin inhibits intracranial glioma growth and negatively regulates in vivo angiogenesis through a CD36-dependent mechanism. Cancer Res. 2009;69:1212–1220. doi: 10.1158/0008-5472.CAN-08-1166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Nishimori H, Shiratsuchi T, Urano T, Kimura Y, Kiyono K, Tatsumi K, Yoshida S, Ono M, Kuwano M, Nakamura Y, Tokino T. A novel brain-specific p53-target gene, BAI1, containing thrombospondin type 1 repeats inhibits experimental angiogenesis. Oncogene. 1997;15:2145–2150. doi: 10.1038/sj.onc.1201542. [DOI] [PubMed] [Google Scholar]
  • 38.Fuxreiter M, Tompa P. Fuzzy complexes: a more stochastic view of protein function. Adv Exp Med Biol. 2012;725:1–14. doi: 10.1007/978-1-4614-0659-4_1. [DOI] [PubMed] [Google Scholar]
  • 39.Klenotic PA, Huang P, Palomo J, Kaur B, Van Meir EG, Vogelbaum MA, Febbraio M, Gladson CL, Silverstein RL. Histidine-rich glycoprotein modulates the anti-angiogenic effects of vasculostatin. Am J Pathol. 2010;176:2039–2050. doi: 10.2353/ajpath.2010.090782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Zhang X, Galardi E, Duquette M, Delic M, Lawler J, Parangi S. Antiangiogenic treatment with the three thrombospondin-1 type 1 repeats recombinant protein in an orthotopic human pancreatic cancer model. Clin Cancer Res. 2005;11:2337–2344. doi: 10.1158/1078-0432.CCR-04-1900. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

3
NIHMS485430-supplement-3.pdf (9.1MB, pdf)
4
NIHMS485430-supplement-4.pdf (184.8KB, pdf)

RESOURCES