GTC
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE ADVANCED SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Genes to Cells (2005) 10, 929-939. doi:10.1111/j.1365-2443.2005.00890.x
© 2005 Blackwell Publishing or its licensors
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stahl, S.
Right arrow Articles by Felbor, U.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stahl, S.
Right arrow Articles by Felbor, U.

Endostatin phenylalanines 31 and 34 define a receptor binding site

Sonja Stahl1, Sabine Gaetzner1, Thomas D. Mueller2 and Ute Felbor1,*

1 Department of Human Genetics, and 2 Department of Physiological Chemistry II, University of Würzburg, Biozentrum, Am Hubland, D-97074 Würzburg, Germany


   Abstract
Top
 Abstract
Introduction
 Results
 Discussion
 Experimental procedures
 References
 
Endostatin has achieved much attention as a naturally occurring inhibitor of angiogenesis and tumor growth. Endostatin is derived from collagen XVIII's C-terminal domain and deleted or truncated in most patients suffering from Knobloch syndrome blindness. To evaluate the functional significance of two surface-exposed hydrophobic phenylalanines at positions 31 and 34 of endostatin and two human sequence alterations within endostatin, A48T and D104N, we applied the alkaline phosphatase fusion protein method. Replacement of F31 and F34 with alanines led to complete loss of characteristic in situ binding while heparin binding remained intact. In contrast, a non-heparin binding alkaline phosphatase-tagged human endostatin lacking R27 and R139 bound to specific tissue structures. The two Knobloch syndrome-associated endostatin sequence variants did not result in altered in situ binding to murine embryonal tissues, human endothelial cells, heparin and immobilized laminin. However, expression of the endostatin mutant A48T was significantly reduced. This observation may be explained by a lower folding efficiency due to the structural constraints of A48 residing in the hydrophobic core. Our data suggest that residues F31 and F34 form a putative receptor binding site acting independently from heparan sulfate binding and that the A48T mutation destabilizes the endostatin molecule.


   Introduction
Top
 Abstract
 Introduction
Results
 Discussion
 Experimental procedures
 References
 
Knobloch syndrome (MIM# 267750 [OMIM] ) is an autosomal recessive disorder characterized by an early onset progressive generalized eye disease and a congenital occipital encephalocele (Knobloch & Layer 1971; Czeizel et al. 1992; Seaver et al. 1993; Passos-Bueno et al. 1994; Wilson et al. 1998; Sniderman et al. 2000). High myopia and vitreoretinal degeneration are the main features of the ocular defect in Knobloch syndrome and lead to bilateral blindness during childhood in most cases. The cranial defect can be corrected surgically within the first year of life. It remains unclear whether other singular abnormalities such as lung hypoplasia, cardiac dextroversion, neuronal migration defects, dilatation of brain ventricles, unilateral duplicated renal collecting systems and generalized hyperextensibility of joints belong to the disease spectrum of Knobloch syndrome.

Knobloch syndrome is caused by homozygous or compound heterozygous mutations in the COL18A1 gene on 21q22.3 (Sertiéet al. 2000; Suzuki et al. 2002; Kliemann et al. 2003; Menzel et al. 2004). Most mutations result in premature stop codons that affect all three alternatively spliced COL18A1 isoforms. A homozygous splice site mutation that exclusively affects collagen XVIII's shortest isoform has been associated with a less severe ocular pathology and loss of vision beyond 20 years of age in a large consanguineous Brazilian kindred (Sertiéet al. 2000). Consistent with the human phenotype, collagen XVIII knockout mice show complex ocular abnormalities including delayed regression of hyaloid arteries and abnormal outgrowth of retinal vessels during the first weeks of life (Fukai et al. 2002; Marneros & Olsen 2003; Ylikärppäet al. 2003; Marneros et al. 2004). On a specific C57BL background, collagen XVIII knockout mice also develop hydrocephalus, dilatation of brain ventricles, and markedly broadened basement membranes in atrioventricular valves of the heart and in the kidney tubules with elevated serum creatinine levels (Utriainen et al. 2004). This suggests that the absence of a functional collagen XVIII molecule can cause other phenotypic alterations depending on genetic background.

Collagens XVIII and XV are non-fibrillar basement membrane molecules characterized by multiple interruptions in their central triple helical region and a unique highly homologous C-terminal domain (Oh et al. 1994a; Rehn & Pihlajaniemi 1994). This globular C-terminal 20 kDa endostatin domain of collagen XVIII is separated from a 50 residue association domain responsible for homotrimerization of collagen {alpha}1(XVIII) chains by a protease-sensitive hinge region (Sasaki et al. 1998). After proteolytic cleavage, endostatin is a potent inhibitor of angiogenesis and tumor growth (O’Reilly et al. 1997; Felbor et al. 2000). However, its receptors and mechanisms of action are still insufficiently characterized. As far as endostatin's physiological role within collagen XVIII is concerned, we recently suggested that endostatin harbors a prominent tissue-binding site for collagen XVIII, based on in situ-binding experiments to murine embryonal eye tissues (Rychkova et al. 2005). This notion agrees with the observation that deletion of the entire C-terminal non-collagenous domain of collagen XVIII still results in a stable truncated product in C. elegans (Ackley et al. 2001). Six of nine known independently arisen COL18A1 mutations are located in exons 35, 36, 40 and 41, leading to a premature stop just before or within the C-terminal endostatin region encoded by exons 41–43 (Suzuki et al. 2002; Menzel et al. 2004). Assuming the importance of endostatin for tissue binding, these mutations would result in absent binding of collagen XVIII to vascular mesenchyme in ocular tissues and endothelial and epithelial basement membranes.

The only two known nucleotide substitutions associated with early-onset progressive Knobloch syndrome are also located within collagen XVIII's endostatin domain: two affected sibs have been reported with a homozygous replacement of a conserved alanine to threonine at position 48 within the endostatin domain (Kliemann et al. 2003). This A48T endostatin mutation corresponds to A179T in the non-collagenous domain of collagen XVIII, a terminology used by Kliemann et al. (2003). In addition, a compound heterozygous D104N endostatin polymorphism (dbSNP/rs12483377) was reported in two further Knobloch sibs (Menzel et al. 2004) but remains of unclear functional relevance (Antonarakis et al. 2005; Suzuki et al. 2005). While this particular variant has also been associated with increased risk for prostate cancer (Iughetti et al. 2001) recent in vitro angiogenesis studies suggest that the D104N endostatin variant is not functionally relevant on its own (Macpherson et al. 2004).

In the present study, we generated a systematic set of human alkaline phosphatase (AP) endostatin fusion proteins in order to analyze in situ binding of human endostatin mutants. We first show that the human endostatin domain of collagen XVIII binds to very specific structures such as the lense capsule in the eye which might suggest a lenticular contribution to the pathogenesis of Knobloch syndrome's myopia. Furthermore, we identify amino acid residues responsible for tissue binding and provide evidence that the recently reported human A48T amino acid substitution within endostatin is of functional significance.


   Results
Top
 Abstract
 Introduction
 Results
Discussion
 Experimental procedures
 References
 
Complementary to expression pattern analyses based on RNA in situ hybridization and immunohistochemistry, dimeric APtags can be used to localize, characterize, and identify functionally active molecules that are capable of ligand-receptor binding (Flanagan et al. 2000). The AP-human endostatin (AP-hES) fusion proteins generated in this study are comprised of a heat-stable, secreted human placental isozyme of AP at the N-terminus and human endostatin at the C-terminus. Endostatin was fused to the C-terminus of AP because endostatin is the C-terminal domain of collagen XVIII. Two hydrophobic phenylalanines at positions 31 and 34 which are unusually exposed in the crystal structure of both murine and human endostatins were replaced by alanines: AP-hESF31/34A. In addition, the A48T replacement reported in two Knobloch sibs and the D104N endostatin polymorphism were introduced into AP-hES: AP-hESA48T and AP-hESD104N (for location of the above residues see Fig. 1A,B). Furthermore, two arginine residues known to be responsible for nonspecific binding of endostatin to heparan sulfate-containing proteoglycans on cell surfaces (Sasaki et al. 1999) were changed to alanines (AP-hESR27/139A). Since the murine equivalent of the arginine/alanine mutant is known to show reduced background staining on tissue sections after longer incubation times (Rychkova et al. 2005), the A48T replacement and the D104N endostatin polymorphism were combined with the arginine double substitutions to generate triple mutant proteins: AP-hESR27/139A/A48T and AP-hESR27/139A/D104N. Finally, unfused AP was used as a negative control.



View larger version (63K):
[in this window]
[in a new window]
  		Figure 1  Stereo view of a ribbon (A) and the respective molecular surface representation (B) of human endostatin (PDB entry code 1BNL [PDB] ). (A) The side chains of residues R27, F31, F34, A48, D104, and R139 investigated in this study are shown as stick models and labeled accordingly. Ribbon and surface representation are oriented identically; the latter clearly shows that A48 is not accessible at the surface and hidden inside the hydrophobic core.

 
Protein production by transiently transfected 293T cells was monitored using quantitative AP activity measurement and Western blot analyses. Apart from AP-hESA48T and AP-hESR27/139A/A48T, all fusion protein-containing supernatants and control AP had AP activities comparable to those obtained from conditioned media containing murine endostatins fused to AP (AP-mES, AP-mESR158/270A, and AP-mESXV) which had been previously generated (Rychkova et al. 2005). Interestingly, cells expressing AP-hESA48T and AP- hESR27/139A/A48T repeatedly produced approximately 40% less (e.g. mean of 1.17 µmol/mL for AP-hESA48T instead of 2.023 µmol/mL for AP-hES, data not shown). The control supernatant of non-transfected 293T cells showed only background activity (data not shown). For all subsequent experiments, the supernatants were diluted to identical volume activities. Western blot analyses demonstrated that a polyclonal antibody against secreted human placental alkaline phosphatase reacted with AP fusion proteins of the expected size in supernatants of transfected cells whereas the supernatant of non-transfected cells did not react with this antibody (data not shown). Fusion proteins were also detected with a rabbit polyclonal anti-human endostatin antibody (Fig. 2A) and quantified by Western blot analyses with known amounts of recombinant human endostatin. Ten microliters conditioned supernatants contained approximately 20 ng of fusion proteins (Fig. 2B).



View larger version (33K):
[in this window]
[in a new window]
  		Figure 2  Representative Western blots probed with a polyclonal anti-human endostatin antibody. (A) All fusion proteins (arrow head) migrated at the expected molecular weight on 5–15% gradient SDS-PAGE gels (AP, alkaline phosphatase, 67 kDa; ES, endostatin, 20–22 kDa; h, human). Note that the antibody detects an endogenous endostatin-containing C-terminal fragment of collagen XVIII (*) indicating that 293T cells produce and process collagen XVIII. (B) Quantification of AP fusion proteins by Western blot analysis using known amounts of recombinant purified endostatin.

 
The equivalence of the highly conserved murine and human AP-endostatin fusion proteins was analyzed using in situ staining of 14-day-old (E14.5) murine embryonal eyes because murine endostatin as well as the murine non-heparin binding endostatin mutant bind particularly well to the tunica vasculosa lentis (Rychkova et al. 2005). Both endostatin affinity probes, AP-mES (data not shown) and AP-hES (Fig. 3A), stained the lense capsule equally well. In contrast, the AP-hESF31/34A fusion protein completely lacked characteristic staining of the lense capsule (Fig. 3B). The mutant AP-hESA48T, AP-hESD104N and AP-hESR27/139A probes resulted in comparable staining patterns (Fig. 3C–E). Control AP did not stain the eye section (Fig. 3F). These results were confirmed by in situ labeling of murine E14.5 sagittal sections (representative examples in Fig. 3G–I). Thus, stainings of all major tissues demonstrated that the observations obtained from ocular tissues can be generalized to endothelial and epithelial basement membranes.



View larger version (93K):
[in this window]
[in a new window]
  		Figure 3  In situ staining of AP fusion proteins on E14.5 murine embryonal eyes and sagittal sections. (A) AP-hES, (C) AP-hESA48T, (D) AP-hESD104N and (E) AP-hESR27/139A stained the developing lense capsule equally well. (B) AP-hESF31/34A did not stain the lense capsule and (F) AP alone was also negative (m, murine; h, human; l, lense; r, retina; tvl, tunica vasculosa lentis). When compared to (G) AP-hES, (H) AP-hESF31/34A did not stain general basement membranes. AP-hESF31/34A gave only slight back-ground staining when compared to (I) unfused AP. This unspecific staining increased upon longer incubation times while AP stayed negative (data not shown) (h, heart; k, definitive kidney; ag, adrenal gland; lu, lung; lv, liver). All sections were incubated with supernatants containing AP fusion proteins for 90 min which is sufficient to reach saturation of binding partners and with the NBT/BCIP substrate for 15 min (eye sections) or 1 h (sagittal sections) (Rychkova et al. 2005).

 
For quantitation of AP fusion protein binding, cultured human dermal microvascular endothelial cells (HMEC-1) were incubated with AP affinity probes for 90 min, washed, lyzed, and assayed for bound AP activity. AP-mES showed stronger binding to HMEC-1 than AP-hES (Fig. 4A). The human A48T and D104N endostatin mutants did not show a statistically significant reduction in affinity to HMEC-1 when compared to AP-hES. In contrast, AP-hESF31/F34h did not bind to HMEC-1. The non-heparin/heparan sulfate binding mutants showed reduced cell binding (data not shown) indicating either that these cells do not contain a significant amount of high affinity receptors or that heparan sulfates are a prerequisite for endostatin binding to these cells.



View larger version (18K):
[in this window]
[in a new window]
  		Figure 4  Quantitative analysis of AP fusion protein binding to the surface of cultured human microvascular endothelial cells (HMEC-1), heparin, and laminin. (A) Human fusion protein binding to HMEC-1 was weaker than murine. Binding affinities of the human A48T endostatin mutant and the D104N polymorphism were only slightly reduced while AP-hESF31/34A showed a statistically significant loss of binding (P = 0.0077). (B) Binding of AP-hESF31/34A, AP-hESA48T and AP-hESD104N to heparin affinity coloumns is not altered. AP-hES, AP-hESF31/34A, AP-hESA48T and AP-hESD104N showed identical profiles and started to elute at 0.38 M NaCl. AP-hESR27/139A was mainly found in the unbound wash through (data not shown) and retained only residual affinity at 0.18–0.26 M NaCl probably due to remaining arginine residues in the core protein. (C) Solid phase assays revealed strikingly reduced binding of AP-hESF31/34A to immobilized laminin when compared to AP-mES, AP-hES, AP-hESA48T and AP-hESD104N.

 
Heparin/heparan sulfate affinity is a key feature of endostatin (Gitay-Goren et al. 1992; Sasaki et al. 1999) and many other growth factors such as vascular endothelial growth factor (Gitay-Goren et al. 1992; Sasaki et al. 1999). Therefore, we analyzed whether the human amino acid changes would lead to reduced/enhanced binding to heparin using affinity chromatography (Fig. 4B). AP-hES, AP-hESF31/34A, AP-hESA48T, and AP-hESD104N all bound to HiTrapTM heparin affinity columns (Pharmacia) and required 0.38 M NaCl for displacement. Thus, these amino acid changes do not influence endostatin's ability to bind to heparin. As expected, the non-heparin/heparan sulfate binding mutant AP-hESR27/139A had only residual affinity to the heparin affinity resin. The affinity of AP-hESF31/34A to heparin provides a possible explanation for the observed unspecific background staining in tissue sections (Fig. 3H).

Collagen XVIII has a polarized location in basement membranes with its C-terminal endostatin domain embedded in the basement membrane facing the endothelial/epithelial cell and the N-terminal non-collagenous domain located at the basement membrane–matrix interface (Fukai et al. 2002; Marneros & Olsen 2003; Ylikärppäet al. 2003; Marneros et al. 2004). Recently, a two receptor model for endostatin binding was proposed. In an in vitro Matrigel tube formation model system, oligomeric endostatin bound to heparan sulfates on cell surfaces and to laminin in the basement membrane (Javaherian et al. 2002). Binding of the immobilized trimeric-non-collagenous domain 1 of collagen XVIII to the extracellular matrix molecule laminin-1 had also been demonstrated in solid-phase assays (Sasaki et al. 1998). In these assays, monomeric endostatin was 100-fold less active. Since the APtag is dimeric and expected to produce a pair of ligand moieties facing away from the tag in the same direction (Flanagan & Cheng 2000), binding of our AP-endostatin fusion proteins to murine laminin-1 was analyzed in solid-phase assays. The AP-mES, AP-hES, AP-hESA48T, and AP-hESD104N probes showed similar affinity to laminin (Fig. 4C) which correlates with the results obtained from in situ labeling. Similarly, absent binding of AP-hESF31/34A to laminin reflected the results obtained from in situ analyses of murine tissues and human cells. As has been described earlier (Javaherian et al. 2002), the non-heparin/heparan sulfate binding mutant AP-hESR27/139A showed significantly reduced binding to laminin. Interactions of AP-mES and AP-hES with immobilized human {alpha}5ß1 integrin, another potential endostatin receptor (Rehn et al. 2001; Sudhakar et al. 2003; Wickström et al. 2003; Wickström et al. 2004), were negligible when compared to laminin and close to control AP (below 0.06 OD405nm/10 min, data not shown).


   Discussion
Top
 Abstract
 Introduction
 Results
 Discussion
Experimental procedures
 References
 
While endostatin's prominent affinity for heparan sulfates appears to be required for anti-angiogenic activity (reviewed in Olsson et al. 2004; Gaetzner et al. 2005), it is not essential for tissue binding (Rychkova et al. 2005). We here demonstrate that the two surface-exposed, highly conserved phenylalanines F31 and F34 are indispensable for in situ binding to murine tissues, human endothelial cells, and immobilized laminin. A dimeric AP-endostatin fusion protein in which F31 and F34 were replaced with alanines bound to heparin but lacked specific in situ binding in agreement with loss of anti-migratory activity of the murine F162/165 A endostatin mutant on human umbilical vein endothelial cells (Karumanchi et al. 2001). F31 and F34 are highly conserved in vertebrates (Fig. 5). The crystal structures of endostatins derived from collagens XV and XVIII revealed that these two phenylalanine residues are in close proximity on the {alpha}1 helix at the center of an absolutely conserved patch (Ding et al. 1998; Hohenester et al. 1998; Sasaki et al. 2000). A comparison of different endostatin domains (PFAM database) from various species shows that the degree of conservation of solvent exposed residues is not randomly distributed with respect to their location on the endostatin molecule (Fig. 5A). The ‘front’ epitope (Fig. 5B, left panel) reveals ten highly conserved residues located within a 10Å radius with the two phenylalanines 31 and 34 in the center. The surface located on the ‘back’ harbors only very few residues which exhibit a similar degree of conservation (Fig. 5B, middle and right panel). In addition, several of those residues are either important for structural integrity (proline and glycine residues) or only partly accessible at the surface (A85, D107, L109, K117, L157) and their conservation might therefore be required for structure maintenance. Since conservation of solvent-exposed phenylalanines throughout evolution is rather unusual (Fig. 5A,B, left panel), this indicates that the epitope centered around the two phenylalanines 31 and 34 might represent an interface for protein-protein interaction and a putative endostatin high affinity receptor binding site (Hohenester et al. 1998). Since we could show that heparin binding is not affected by these phenylalanine to alanine substitutions, this epitope must represent a binding site to a new, so far unidentified receptor. However, the binding epitopes for heparan sulfate and the unidentified receptor might partially overlap due to their close proximity. Interestingly, in human endostatin crystals lattice contacts involve F31 and F34 from each monomer but it remained unclear whether these hydrophobic interactions occur in vivo (Ding et al. 1998). The crystal lattice contacts involving the two conserved phenylalanines are not only different between crystals of human and murine endostatin but also for crystals of murine endostatin with a different spacegroup (Hohenester et al. 2000). Therefore, these contacts rather display the ‘stickiness’ of the hydrophobic surface patch than dimerization through phenylalanines.



View larger version (84K):
[in this window]
[in a new window]
  		Figure 5  (A) Sequence alignment of a selection of endostatin domains from the ‘full alignment’ of the PFAM database (http://www.sanger.ac.uk/Software/Pfam/) and the NCBI protein database prepared with the software Jalview (Clamp et al. 2004). The Taylor color coding was used for the alignment, the color intensities are modulated by the degree of conservation, the color intensities shown were produced with a scale of 75 of the conservation tool in Jalview. Secondary structure elements are shown as arrows and helices according to Ding et al. (1998). Residues investigated in this study are marked by asterisks. (B) Surface representation of human endostatin (PDB entry code 1BNL [PDB] ) color coded according to the degree of conservation shown in (A). All residues which are still colored at a conservation level of 90 (Jalview tool conservation) are shown in red, residues marked at a level of 75 are shown in orange, whereas residues shown in yellow are colored only below a level of 75 but above 50. Therefore, residues in red mark the highest degree of conservation, residues in yellow the lowest. The orientation of the molecule shown in the left panel is identical with that in Fig. 1. The orientation of the molecule displayed in the middle and right panel is rotated by 120° around the y-axis, respectively. The figure clearly shows that the conservation of surface-accessible residues is not randomly distributed over the molecule's surface. Instead there is a clear clustering with the phenylalanines 31 and 34 in the center of a putative protein binding epitope.

 
We also present the first functional evaluation of the homozygous A48T amino acid substitution found in two sisters affected with Knobloch syndrome but not in 100 control individuals (Kliemann et al. 2003). While in situ binding affinities were not altered, the production of AP-hESA48T was strikingly reduced when compared to wild-type endostatins. Since A48 is not accessible on the endostatin surface (Fig. 1A,B), we conclude that the effect must be indirect. Analyzing the immediate environment of A48 reveals that its side chain is pointing into the interior of endostatin's hydrophobic core and is completely shielded from the solvent. Due to the tight packing of A48 by its direct neighboring residues I26, A29, D30 and L50 space for larger amino acid side chains is limited. Replacing the alanine with the more hydrophilic threonine side chain might therefore cause a lower folding efficiency. In the hydrophobic environment the polar hydroxyl group of threonine must be balanced by a hydrogen bond. Possible hydrogen bond donors/acceptors close in space are only the side chain carboxyl group or the main chain amide of D30. This stringent restraint might therefore lead to an increased amount of misfolded protein consistent with our findings that the expression rate of AP-hESA48T is reduced while the purified protein shows otherwise unaltered binding properties. A48's importance for the structural maintenance can also be deduced from the absolute conservation of the alanine residue at this position throughout all 40 endostatin domains in the PFAM database (http://www.sanger.ac.uk/Software/Pfam/). Although Kliemann et al. (2003) did only screen 25 of the 43 collagen XVIII exons by SSCP, it seems likely that A48T is a pathological mutation.

In contrast, our in situ stainings of tissues and cells, the heparin affinity chromatographies, and solid-phase assays did not provide evidence that the D104N amino acid substitution is of major pathological and clinical significance. Two recent studies have addressed functional aspects of D104N. Menzel et al. (2004) found impaired affinity of endostatin D104N for laminin by immunoprecipitation and Western blotting. However, this result might have been due to a different experimental approach in which Flag-tagged endostatin dimers were created via biotinylated anti-Flag IgG. Notably, Menzel et al. (2004) did not observe a reduced inhibitory activity of the D104N variant on migration of human dermal microvascular endothelial cells. In addition, it was shown that recombinant human D104 and D104N endostatins inhibit human umbilical vein endothelial cell tube formation equally well (Macpherson et al. 2004). The side chain of D104 is fully exposed to solvent and not involved in any non-covalent interactions that are required for structural integrity of the endostatin molecule (Fig. 1A,B). Replacement by the isosterical asparagine should therefore be without any consequences for the folding of the molecule. Taken together, the previous activity assays, our binding studies and the structural data do not confirm the hypothesis that the exchange of aspartic acid to asparagine at position 104 in endostatin (D104N) would directly lead to impaired anti-angiogenic activity and decreased affinity to other molecules (Iughetti et al. 2001).

As already pointed out by Menzel et al. (2004), it is possible that the two Hungarian Knobloch sibs carrying a heterozygous frameshift mutation and the endostatin D104N polymorphism may have an additional, yet undetected pathogenic mutation in e.g. regulatory elements or a large deletion. This possibility is supported by a recent report on a single unaffected mother of a Knobloch patient who carried a frameshift insertion in trans with the D104N polymorphism (Suzuki et al. 2005). Furthermore, the D104N variant is polymorphic in various populations with a frequency of 5.6–15% (Iughetti et al. 2001; Macpherson et al. 2004; Menzel et al. 2004), and homozygous individuals are healthy (Suzuki et al. 2005). Nevertheless it cannot be excluded that compound heterozygotes bearing one null allele in combination with the D104N allele show a reduced penetrance of the phenotype (Antonarakis et al. 2005).

Regarding predisposition to prostate cancer reported for the endostatin D104N polymorphism (Iughetti et al. 2001), no statistically significant association between the frequency of endostatin D104N and the incidence of androgen independent prostate cancer or survival was found in a larger case control group (Macpherson et al. 2004) and in tissues from prostate cancer patients (Li et al. 2005). However, the replication of association studies may not succeed due to different genetic backgrounds of the populations used and the heterogeneity of the disease studied. Therefore, it remains conceivable that the endostatin D104N polymorphism is in linkage disequilibrium with a collagen XVIII mutation of functional relevance.


   Experimental procedures
Top
 Abstract
 Introduction
 Results
 Discussion
 Experimental procedures
References
 
Construction and expression of alkaline phosphatase fusion proteins

To generate human AP-endostatin fusion proteins, the human endostatin sequence was amplified from cDNA clone {alpha}1(XVIII) pNF18-2 (Oh et al. 1994b), mutated in pBluescript II SK(+) (Stratagene) using the QuickChangeTM Site-Directed Mutagenesis kit (Stratagene), and subcloned into the expression vector pAPtag-4 (obtained from D.A. Feldheim and J.G. Flanagan, Harvard Medical School) using oligonucleotide primers AP-hES5' (5'-GGT TCC GGA CAC AGC CAC CGC GAC TTC CAG-3') and AP-hES3' (5'-ATG CTC GAG CTA CTT GGA GGC AGT CAT GAA-3') essentially as described (Rychkova et al. 2005). 293T cells (human embryonic kidney cells, a gift from D.A. Feldheim and J.G. Flanagan) plated at 80% confluence on 150 mm tissue culture plates were transiently transfected with 12 µg of fusion plasmid DNA using FuGENE 6 transfection reagent (Roche). Twenty-four hours after transfection, the medium (DMEM with glutamax-I (Gibco), 10% fetal bovine serum, 1% penicilline-streptomycine) was replaced. Conditioned supernatants from transfected and non-transfected cells were collected after an additional 48–72 h, centrifuged at 1000 r.p.m. (Eppendorf rotor A-4-44), filtered through a 0.45 µm filter (Schleicher & Schuell), buffered with 10 mM HEPES, 0.05% NaN3, pH 7.0, and stored at 4 °C for immediate use or at –80 °C for longterm usage.

Determination of AP activity

One hundred microliters supernatants were heat-inactivated for 10 min at 65 °C to inhibit endogenous phosphatases. After centrifugation at 14000 r.p.m. (Eppendorf rotor F45-30-11), 20 µL were mixed with 380 µL HBAH buffer (150 mM NaCl, 20 mM HEPES, pH 7.0, 0.5 mg/mL bovine serum albumin, 0.1% NaN3) and 400 µL 2x AP substrate buffer (2 M diethanolamine, 1 mM MgCl2, 18 mM p-nitrophenyl phosphate (AppliChem), pH 9.8), and incubated at room temperature. Absorbance at 405 nm was read at 30 s intervals for 10 min in a spectrophotometer. After AP activity measurement, the supernatants were diluted to obtain the same activity in each probe. In subsequent experiments, specific activities were compared to wild-type endostatin. AP fusion proteins were not affinity purified, quantitated, and titrated to perform exact enzyme kinetics.

SDS-PAGE/Western blotting

Fifteen microliters of conditioned supernatants were loaded on to 5–15% SDS-PAGE gradient gels, run at 50 V for 18 h and wetblotted on to nitrocellulose (Protran; Schleicher & Schuell) using a TE42 Transphor transfer unit (Amersham Biosciences) 1.5 A for 1 h at 4 °C. The blots were incubated with a rabbit polyclonal antibody against secreted human placental alkaline phosphatase (1 : 2000, WAK-Chemie) or with a rabbit polyclonal anti-human endostatin antibody (1 : 2000, Cytimmune Sciences Inc.) followed by a horseradish peroxidase-conjugated anti-rabbit IgG (1 : 3000, Santa Cruz) according to standard procedures. The signals were visualized using enhanced chemiluminescence (Perkin Elmer/NEN).

Staining of tissue sections

Timed-mated NMRI mice were ordered from Harlan Winkelmann. E14.5 embryos were dissected, fixed in 4% paraformaldehyde (in PBS) at 4 °C overnight, transferred to 20% sucrose (in PBS) at 4 °C on a shaker for one day, and frozen in OCT embedding medium (Tissue-Tek). AP-staining of 10 µm cryosections thaw-mounted on to Polysine slides (Menzel Gläser) was essentially performed as described in Flanagan et al. (2000) and Rychkova et al. (2005).

Quantitative measurement of AP-endostatin binding to cell surfaces

Human dermal microvascular endothelial cells (HMEC-1 cell line kindly provided by the Centers for Disease Control and Prevention, Atlanta, GA, USA) were plated into six-well tissue culture plates and cultured for one day in Endothelial Basal Medium (Clonetics) supplemented with 10 ng/mL human epidermal growth factor (Clonetics), 1 µg/mL hydrocortisone (Clonetics), and 10% fetal bovine serum (Linaris). Confluent cells were washed once with cold HBAH buffer, incubated with 1 mL fusion protein-containing supernatants for 90 min at room temperature, washed with cold HBAH five times for 5 min and lyzed with 300 µL 1% Triton X-100, 10 mM Tris-HCl, pH 8.0. After collection of the lysates, the plates were rinsed again with an additional 200 µL of lysate buffer. The pooled lysates were vortexed, incubated at room temperature for 5 min, heat-inactivated at 65 °C for 10 min, placed on ice, and supplemented with an equal amount (400 µL) of 2x AP substrate buffer to measure the AP activity after 30 min as described above. The experiments were performed in triplicate. A detailed protocol for this procedure can be found in Flanagan & Cheng 2000).

Heparin affinity chromatography

Sixty milliliters of filtered supernatants adjusted to equal activities were applied to 5 mL HiTrapTM heparin affinity coloumns (Pharmacia) equilibrated in 0.05 M Tris-HCl, pH 7.4. The protein samples were eluted with a linear gradient of 0–1 M NaCl. The content of the individual fractions was determined by AP activity measurements and Western blotting with the anti-AP and anti-endostatin antibodies.

Solid-phase assays

Ninety-six-well flat bottom plates (Greiner) were coated with 1 µg purified murine laminin-1 (BD Biosciences) or 0.5 µg purified human {alpha}5ß1 integrin (Chemicon) per well at 4 °C overnight. After three washes with cold HBAH buffer, immobilized laminin was incubated with 100 µL AP affinity probes for 90 min at room temperature. The probes were supplemented with 2 mM CaCl2, 1 mM MgCl2, and 1 mM MnCl2 in solid phase assays with {alpha}5ß1 integrin. Unbound protein was removed by three washes with HBAH buffer. One hundred microliters 2x AP substrate solution (2 M diethanolamine, 1 mM MgCl2, 18 mMp-nitrophenyl phosphate (AppliChem), pH 9.8) was added to 100 µL HBAH buffer per well. Absorbance was measured at 405 nm after 10 min using a Dynatech MR5000 plate reader. Coated wells containing cell culture medium were used for zero adjustment. The experiments were performed in triplicate.


   Acknowledgements
 
We thank Drs. B.R. Olsen and N. Fukai for cDNA clone {alpha}1(XVIII) pNF18-2, Drs. D.A. Feldheim and J.G. Flanagan for the expression vector pAPtag-4, Drs. E. Ades, F.J. Candal and T. Lawley for the HMEC-1 cell line, T. Neumann for her contribution to the biochemical experiments, and Dr. E. Conzelmann for helpful discussions. This work was supported by an Emmy Noether-grant from the Deutsche Forschungsgemeinschaft (Fe 432/6-4).


   Footnotes
 
Communicated by: Yo-ichi Nabeshima

* Correspondence: E-mail: felbor{at}biozentrum.uni-wuerzburg.de


   References
Top
 Abstract
 Introduction
 Results
 Discussion
 Experimental procedures
 References
 
Ackley, B.D., Crew, J.R., Elamaa, H., Pihlajaniemi, T., Kuo, C.J. & Kramer, J.M. (2001) The NC1/endostatin domain of Caenorhabditis elegans type XVIII collagen affects cell migration and axon guidance. J. Cell Biol. 152, 1219–1232.[Abstract/Free Full Text]

Antonarakis, S.E., Reymond, A., Menzel, O., et al. (2005) A response to Suzuki et al. ‘How pathogenic is the p.D104N/endostatin polymorphic allele of COL18A1 in Knobloch syndrome?’. Hum. Mutat. 25, 316.[CrossRef][Medline]

Clamp, M., Cuff, J., Searle, S.M. & Barton, G.J. (2004) The Jalview Java alignment editor. Bioinformatics 20, 426–427.[Abstract/Free Full Text]

Czeizel, A.E., Goblyos, P., Kustos, G., Mester, E. & Paraicz, E. (1992) The second report of Knobloch syndrome. Am. J. Med. Genet. 42, 777–779.[CrossRef][Medline]

Ding, Y.H., Javaherian, K., Lo, K.M., et al. (1998) Zinc-dependent dimers observed in crystals of human endostatin. Proc. Natl. Acad. Sci. USA 95, 10443–10448.[Abstract/Free Full Text]

Felbor, U., Dreier, L., Bryant, R.A., Ploegh, H.L., Olsen, B.R. & Mothes, W. (2000) Secreted cathepsin L generates endostatin from collagen XVIII. EMBO J. 19, 1187–1194.[CrossRef][Medline]

Flanagan, J.G. & Cheng, H.J. (2000) Alkaline phosphatase fusion proteins for molecular characterization and cloning of receptors and their ligands. Methods Enzymol. 327, 198–210.[Medline]

Flanagan, J.G., Cheng, H.J., Feldheim, D.A., Hattori, M., Lu, Q. & Vanderhaeghen, P. (2000) Alkaline phosphatase fusions of ligands or receptors as in situ probes for staining of cells, tissues, and embryos. Methods Enzymol. 327, 19–35.[Medline]

Fukai, N., Eklund, L., Marneros, A.G., et al. (2002) Lack of collagen XVIII/endostatin results in eye abnormalities. EMBO J. 21, 1535–1544.[CrossRef][Medline]

Gaetzner, S., Deckers, M.M.L., Stahl, S., Löwik, C., Olsen, B.R. & Felbor, U. (2005) Endostatin's heparan sulfate-binding site is essential for inhibition of angiogenesis and enhances in situ binding to capillary-like structures in bone explants. Matrix Biol. 23, 557–561.[CrossRef][Medline]

Gitay-Goren, H., Soker, S., Vlodavsky, I. & Neufeld, G. (1992) The binding of vascular endothelial growth factor to its receptors is dependent on cell surface-associated heparin-like molecules. J. Biol. Chem. 267, 6093–6098.[Abstract/Free Full Text]

Hohenester, E., Sasaki, T., Olsen, B.R. & Timpl, R. (1998) Crystal structure of the angiogenesis inhibitor endostatin at 1.5 Å resolution. EMBO J. 17, 1656–1664.[CrossRef][Medline]

Hohenester, E., Sasaki, T., Mann, K. & Timpl, R. (2000) Variable zinc coordination in endostatin. J. Mol. Biol. 297, 1–6.[CrossRef][Medline]

Iughetti, P., Suzuki, O., Godoi, P.H., et al. (2001) A polymorphism in endostatin, an angiogenesis inhibitor, predisposes for the development of prostatic adenocarcinoma. Cancer Res. 61, 7375–7378.[Abstract/Free Full Text]

Javaherian, K., Park, S.Y., Pickl, W.F., et al. (2002) Laminin modulates morphogenic properties of the collagen XVIII endostatin domain. J. Biol. Chem. 277, 45211–45218.[Abstract/Free Full Text]

Karumanchi, S.A., Jha, V., Ramchandran, R., et al. (2001) Cell surface glypicans are low-affinity endostatin receptors. Mol. Cell 7, 811–822.[CrossRef][Medline]

Kliemann, S.E., Waetge, R.T., Suzuki, O.T., Passos-Bueno, M.R. & Rosemberg, S. (2003) Evidence of neuronal migration disorders in Knobloch syndrome: clinical and molecular analysis of two novel families. Am. J. Med. Genet. 119A, 15–19.

Knobloch, W. & Layer, J. (1971) Retinal detachment and encephalocele. J. Pediatric Ophthalmol. 8, 181–184.

Li, H.C., Cai, Q.Y., Shinohara, E.T., et al. (2005) Endostatin polymorphism 4349G/A (D104N) is not associated with aggressiveness of disease in prostate cancer. Dis. Markers 21, 37–41.[Medline]

Macpherson, G.R., Singh, A.S., Bennett, C.L., et al. (2004) Genotyping and functional analysis of the D104N variant of human endostatin. Cancer Biol. Ther. 3, 1298–1303.[Medline]

Marneros, A.G. & Olsen, B.R. (2003) Age-dependent iris abnormalities in collagen XVIII/endostatin deficient mice with similarities to human pigment dispersion syndrome. Invest. Ophthalmol. Vis. Sci. 44, 2367–2372.[Abstract/Free Full Text]

Marneros, A.G., Keene, D.R., Hansen, U., et al. (2004) Collagen XVIII/endostatin is essential for vision and retinal pigment epithelial function. EMBO J. 23, 89–99.[CrossRef][Medline]

Menzel, O., Bekkeheien, R.C., Reymond, A., et al. (2004) Knobloch syndrome: novel mutations in COL18A1, evidence for genetic heterogeneity, and a functionally impaired polymorphism in endostatin. Hum. Mutat. 23, 77–84.[CrossRef][Medline]

Oh, S.P., Kamagata, Y., Muragaki, Y., Timmons, S., Ooshima, A. & Olsen, B.R. (1994a) Isolation and sequencing of cDNAs for proteins with multiple domains of Gly-Xaa-Yaa repeats identify a distinct family of collagenous proteins. Proc. Natl. Acad. Sci. USA 91, 4229–4233.[Abstract/Free Full Text]

Oh, S.P., Warman, M.L., Seldin, M.F., et al. (1994b) Cloning of cDNA and genomic DNA encoding human type XVIII collagen and localization of the alpha 1 (XVIII) collagen gene to mouse chromosome 10 and human chromosome 21. Genomics 19, 494–499.[CrossRef][Medline]

Olsson, A.K., Johansson, I., Akerud, H., et al. (2004) The minimal active domain of endostatin is a heparin-binding motif that mediates inhibition of tumor vascularization. Cancer Res. 64, 9012–9017.[Abstract/Free Full Text]

O'Reilly, M.S., Boehm, T., Shing, Y., et al. (1997) Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 88, 277–285.[CrossRef][Medline]

Passos-Bueno, M.R., Marie, S.K., Monteiro, M., et al. (1994) Knobloch syndrome in a large Brazilian consanguineous family: confirmation of autosomal recessive inheritance. Am. J. Med. Genet. 52, 170–173.[CrossRef][Medline]

Rehn, M. & Pihlajaniemi, T. (1994) Alpha 1 (XVIII), a collagen chain with frequent interruptions in the collagenous sequence, a distinct tissue distribution, and homology with type XV collagen. Proc. Natl. Acad. Sci. USA 91, 4234–4238.[Abstract/Free Full Text]

Rehn, M., Veikkola, T., Kukk-Valdre, E., et al. (2001) Interaction of endostatin with integrins implicated in angiogenesis. Proc. Natl. Acad. Sci. USA 98, 1024–1029.[Abstract/Free Full Text]

Rychkova, N., Stahl, S., Gaetzner, S. & Felbor, U. (2005) Non-heparan sulfate–binding interactions of endostatin/collagen XVIII in murine development. Dev. Dyn. 232, 399–407.[CrossRef][Medline]

Sasaki, T., Fukai, N., Mann, K.G., Göhring, W., Olsen, B.R. & Timpl, R. (1998) Structure, function and tissue forms of the C-terminal globular domain of collagen XVIII containing the angiogenesis inhibitor endostatin. EMBO J. 17, 4249–4256.[CrossRef][Medline]

Sasaki, T., Larsson, H., Kreuger, J., et al. (1999) Structural basis and potential role of heparin/heparan sulfate binding to the angiogenesis inhibitor endostatin. EMBO J. 18, 6240–6248.[CrossRef][Medline]

Sasaki, T., Larsson, H., Tisi, D., Claesson-Welsh, L., Hohenester, E. & Timpl, R. (2000) Endostatins derived from collagens XV and XVIII differ in structural and binding properties, tissue distribution and anti-angiogenic activity. J. Mol. Biol. 301, 1179–1190.[CrossRef][Medline]

Seaver, L.H., Joffe, L., Spark, R.P., Smith, B.L. & Hoyme, H.E. (1993) Congenital scalp defects and vitreoretinal degeneration: redefining the Knobloch syndrome. Am. J. Med. Genet. 46, 203–208.[CrossRef][Medline]

Sertié, A.L., Sossi, V., Camargo, A.A., Zatz, M., Brahe, C. & Passos-Bueno, M.R. (2000) Collagen XVIII, containing an endogenous inhibitor of angiogenesis and tumor growth, plays a critical role in the maintenance of retinal structure and in neural tube closure (Knobloch syndrome). Hum. Mol. Genet. 9, 2051–2058.[Abstract/Free Full Text]

Sniderman, L.C., Koenekoop, R.K., O'Gorman, A.M., et al. (2000) Knobloch syndrome involving midline scalp defect of the frontal region. Am. J. Med. Genet. 90, 146–149.[CrossRef][Medline]

Sudhakar, A., Sugimoto, H., Yang, C., Lively, J., Zeisberg, M. & Kalluri, R. (2003) Human tumstatin and human endostatin exhibit distinct antiangiogenic activities mediated by alpha v beta 3 and alpha 5 beta 1 integrins. Proc. Natl. Acad. Sci. USA 100, 4766–4771.[Abstract/Free Full Text]

Suzuki, O.T., Sertié, A.L., Der Kaloustian, V.M., et al. (2002) Molecular analysis of collagen XVIII reveals novel mutations, presence of a third isoform, and possible genetic heterogeneity in Knobloch syndrome. Am. J. Hum. Genet. 71, 1320–1329.[CrossRef][Medline]

Suzuki, O.T., Bagatini, K., Sertié, A.L. & Passos-Bueno, M. (2005) How pathogenic is the p.D104N/endostatin polymorphic allele of COL18A1 in Knobloch syndrome? Hum. Mutat. 25, 314–315.[CrossRef][Medline]

Utriainen, A., Sormunen, R., Kettunen, M., et al. (2004) Structurally altered basement membranes and hydrocephalus in a type XVIII collagen deficient mouse line. Hum. Mol. Genet. 13, 2089–2099.[Abstract/Free Full Text]

Wickström, S.A., Alitalo, K. & Keski-Oja, J. (2003) Endostatin associates with lipid rafts and induces reorganization of the actin cytoskeleton via down-regulation of RhoA activity. J. Biol. Chem. 278, 37895–37901.[Abstract/Free Full Text]

Wickström, S.A., Alitalo, K. & Keski-Oja, J. (2004) An endostatin-derived peptide interacts with integrins and regulates actin cytoskeleton and migration of endothelial cells. J. Biol. Chem. 279, 20178–20185.[Abstract/Free Full Text]

Wilson, C., Aftimos, S., Pereira, A. & McKay, R. (1998) Report of two sibs with Knobloch syndrome (encephalocoele and viteroretinal degeneration) and other anomalies. Am. J. Med. Genet. 78, 286–290.[CrossRef][Medline]

Ylikärppä, R., Eklund, L., Sormunen, R., et al. (2003) Lack of type XVIII collagen results in anterior ocular defects. FASEB J. 17, 2257–2259.[Abstract/Free Full Text]

Received: 26 April 2005
Accepted: 23 June 2005




This article has been cited by other articles:


Home page
 J. Biol. Chem.Home page
C. Faye, C. Moreau, E. Chautard, R. Jetne, N. Fukai, F. Ruggiero, M. J. Humphries, B. R. Olsen, and S. Ricard-Blum
Molecular Interplay between Endostatin, Integrins, and Heparan Sulfate
J. Biol. Chem., August 14, 2009; 284(33): 22029 - 22040.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stahl, S.
Right arrow Articles by Felbor, U.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stahl, S.
Right arrow Articles by Felbor, U.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE ADVANCED SEARCH TABLE OF CONTENTS