Specific Structural Features of Heparan Sulfate Proteoglycans Potentiate Neuregulin-1 Signaling

Abstract

Neuregulins are a family of growth and differentiation factors that act through activation of cell-surface erbB receptor tyrosine kinases and have essential functions both during development and on the growth of cancer cells. One alternatively spliced neuregulin-1 form has a distinct heparin-binding immunoglobulin-like domain that enables it to adhere to heparan sulfate proteoglycans at key locations during development and substantially potentiates its activity. We examined the structural specificity needed for neuregulin-1-heparin interactions using a gel mobility shift assay together with an assay that measures the ability of specific oligosaccharides to block erbB receptor phosphorylation in L6 muscle cells. Whereas the N-sulfate group of heparin was most important, the 2-O-sulfate and 6-O-sulfate groups also contributed to neuregulin-1 binding in these two assays. Optimal binding to neuregulin-1 required eight or more heparin disaccharides; however, as few as two disaccharides were still able to bind neuregulin-1 to a lesser extent. The physiological importance of this specificity was shown both by chemical and siRNA treatment of cultured muscle cells. Pretreatment of muscle cells with chlorate that blocks all sulfation or with an siRNA that selectively blocks N-sulfation significantly reduced erbB receptor activation by neuregulin-1 but had no effect on the activity of neuregulin-1 that lacks the heparin-binding domain. These results suggest that the regulation of glycosaminoglycan sulfation is an important biological mechanism that can modulate both the localization and potentiation of neuregulin-1 signaling. Neuregulins are a family of growth and differentiation factors that act through activation of cell-surface erbB receptor tyrosine kinases and have essential functions both during development and on the growth of cancer cells. One alternatively spliced neuregulin-1 form has a distinct heparin-binding immunoglobulin-like domain that enables it to adhere to heparan sulfate proteoglycans at key locations during development and substantially potentiates its activity. We examined the structural specificity needed for neuregulin-1-heparin interactions using a gel mobility shift assay together with an assay that measures the ability of specific oligosaccharides to block erbB receptor phosphorylation in L6 muscle cells. Whereas the N-sulfate group of heparin was most important, the 2-O-sulfate and 6-O-sulfate groups also contributed to neuregulin-1 binding in these two assays. Optimal binding to neuregulin-1 required eight or more heparin disaccharides; however, as few as two disaccharides were still able to bind neuregulin-1 to a lesser extent. The physiological importance of this specificity was shown both by chemical and siRNA treatment of cultured muscle cells. Pretreatment of muscle cells with chlorate that blocks all sulfation or with an siRNA that selectively blocks N-sulfation significantly reduced erbB receptor activation by neuregulin-1 but had no effect on the activity of neuregulin-1 that lacks the heparin-binding domain. These results suggest that the regulation of glycosaminoglycan sulfation is an important biological mechanism that can modulate both the localization and potentiation of neuregulin-1 signaling. Heparan sulfate proteoglycans (HSPGs) 1The abbreviations used are: HSPG, heparan sulfate proteoglycan; FGF, fibroblast growth factor; NRG, neuregulin; EGF, epidermal growth factor; Ig, immunoglobulin; IgEGF, form of NRG with both Ig and EGF-like domains; GAG, glycosaminoglycan; p185, tyrosine-phosphorylated erbB receptors; GMSA, gel mobility shift assay; NDST, N-deacetylase/N-sulfotransferase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; siRNA, small interference RNA. 1The abbreviations used are: HSPG, heparan sulfate proteoglycan; FGF, fibroblast growth factor; NRG, neuregulin; EGF, epidermal growth factor; Ig, immunoglobulin; IgEGF, form of NRG with both Ig and EGF-like domains; GAG, glycosaminoglycan; p185, tyrosine-phosphorylated erbB receptors; GMSA, gel mobility shift assay; NDST, N-deacetylase/N-sulfotransferase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; siRNA, small interference RNA. are ubiquitous components of the cell surface and extracellular matrix that are critical for normal development and regulation of cell growth (1Lander A.D. Selleck S.B. J. Cell Biol. 2000; 148: 227-232Crossref PubMed Scopus (223) Google Scholar, 2Gabius H.J. Naturwissenschaften. 2000; 87: 108-121Crossref PubMed Scopus (237) Google Scholar, 3Inatani M. Irie F. Plump A.S. Tessier-Lavigne M. Yamaguchi Y. Science. 2003; 302: 1044-1046Crossref PubMed Scopus (355) Google Scholar). HSPGs have been implicated in a wide range of biological processes including cell adhesion, motility, proliferation, differentiation, tissue morphogenesis, and carcinogenesis (4Turnbull J. Drummond K. Huang Z. Kinnunen T. Ford-Perriss M. Murphy M. Guimond S. Biochem. Soc. Trans. 2003; 31: 343-348Crossref PubMed Scopus (23) Google Scholar, 5Forsberg E. Kjellen L. J. Clin. Investig. 2001; 108: 175-180Crossref PubMed Scopus (166) Google Scholar). Many of these biological processes are mediated through numerous heparin-binding growth and differentiation factors that include fibroblast growth factors (FGFs) (6Deguchi Y. Okutsu H. Okura T. Yamada S. Kimura R. Yuge T. Furukawa A. Morimoto K. Tachikawa M. Ohtsuki S. Hosoya K. Terasaki T. J. Neurochem. 2002; 83: 381-389Crossref PubMed Scopus (56) Google Scholar, 7Izvolsky K.I. Zhong L. Wei L. Yu Q. Nugent M.A. Cardoso W.V. Am. J. Physiol. 2003; 285: L838-L846Google Scholar), platelet-derived growth factor (8Garcia-Olivas R. Hoebeke J. Castel S. Reina M. Fager G. Lustig F. Vilaro S. Histochem. Cell Biol. 2003; 120: 371-382Crossref PubMed Scopus (36) Google Scholar), bone morphogenic proteins (9Irie A. Habuchi H. Kimata K. Sanai Y. Biochem. Biophys. Res. Commun. 2003; 308: 858-865Crossref PubMed Scopus (105) Google Scholar), heparin-binding growth-associated molecule (10Deuel T.F. Zhang N. Yeh H.J. Silos-Santiago I. Wang Z.Y. Arch. Biochem. Biophys. 2002; 397: 162-171Crossref PubMed Scopus (200) Google Scholar), and neuregulins (NRGs) (11Loeb J.A. J. Neurocytol. 2003; 32: 649-664Crossref PubMed Scopus (28) Google Scholar). The NRGs have many functions in the nervous system and cardiac development and in the regulation of cell growth and carcinogenesis (12Falls D.L. Exp. Cell Res. 2003; 284: 14-30Crossref PubMed Scopus (844) Google Scholar). Four NRG genes have been discovered thus far of which NRG1 is the best characterized. The NRG1 gene produces multiple alternatively spliced membrane-bound and soluble isoforms. Soluble isoforms of NRG1 are composed of an epidermal growth factor (EGF)-like domain, which is sufficient for receptor binding and activation of homodimers and heterodimers of erbB2, erbB3, and erbB4, and an immunoglobulin-like domain, NH2-terminal to the EGF-like domain, that is responsible for binding heparin (13Li Q. Loeb J.A. J. Biol. Chem. 2001; 276: 38068-38075Abstract Full Text Full Text PDF PubMed Google Scholar, 14Loeb J.A. Fischbach G.D. J. Cell Biol. 1995; 130: 127-135Crossref PubMed Scopus (114) Google Scholar, 15Meier T. Masciulli F. Moore C. Schoumacher F. Eppenberger U. Denzer A.J. Jones G. Brenner H.R. J. Cell Biol. 1998; 141: 715-726Crossref PubMed Scopus (97) Google Scholar). Interactions with HSPGs appear to direct the distribution of NRG1 in the developing nervous system. For example, HSPGs including agrin become restricted to the basal lamina of neuromuscular synapses coincident with NRG1 during embryonic development in the chick (16Loeb J.A. Khurana T.S. Robbins J.T. Yee A.G. Fischbach G.D. Development. 1999; 126: 781-791Crossref PubMed Google Scholar). Here agrin and NRG1 can work synergistically by forming localized complexes that optimally concentrate acetylcholine receptors at the synapses (17Li Q. Esper R.M. Loeb J. Mol. Cell. Neurosci. 2004; 26: 558-569Crossref PubMed Scopus (15) Google Scholar). This synergy results by both potentiation of NRG1-induced erbB receptor phosphorylation and by producing a sustained signal critical for up-regulating acetylcholine receptor genes (13Li Q. Loeb J.A. J. Biol. Chem. 2001; 276: 38068-38075Abstract Full Text Full Text PDF PubMed Google Scholar). NRG1 accumulation along early motor and sensory axons, radial glia, spinal axonal tracts, and neuroepithelial cells through associations with HSPGs raises other potentially important roles for NRG1 in nervous system development (16Loeb J.A. Khurana T.S. Robbins J.T. Yee A.G. Fischbach G.D. Development. 1999; 126: 781-791Crossref PubMed Google Scholar). Because HSPGs are expressed on the surface of virtually every cell, a key question is how heparin-binding growth factors preferentially associate with HSPGs on some cells but not others. Recent evidence suggests that this specificity may be encoded by the diversity generated by differences in sugar composition and sulfation pattern of the glycosaminoglycan (GAG) chains (2Gabius H.J. Naturwissenschaften. 2000; 87: 108-121Crossref PubMed Scopus (237) Google Scholar, 18Capila I. Linhardt R.J. Angew. Chem. Int. Ed. Engl. 2002; 41: 391-412Crossref PubMed Scopus (1531) Google Scholar). HSPGs are composed of repeating disaccharides of glucuronic acid or iduronic acid and N-acetyl or N-sulfoglucosamine with potential sulfation sites located on the amino group or positions 2, 3, and 6 of each sugar molecule (19Sugahara K. Kitagawa H. IUBMB Life. 2002; 54: 163-175Crossref PubMed Scopus (216) Google Scholar). Both FGF1 and FGF2 require N-sulfated pentasaccharide sequences for optimal binding but differ in the requirements for O-sulfation (20Kreuger J. Salmivirta M. Sturiale L. Gimenez-Gallego G. Lindahl U. J. Biol. Chem. 2001; 276: 30744-30752Abstract Full Text Full Text PDF PubMed Scopus (201) Google Scholar). Hepatocyte growth factor, platelet-derived growth factor, and lipoprotein lipase all depend on the presence of one or more N-acetylglucosamine 6-O-sulfate groups (21Lyon M. Gallagher J.T. Biochem. Soc. Trans. 1994; 22: 365-370Crossref PubMed Scopus (27) Google Scholar, 22Feyzi E. Lustig F. Fager G. Spillmann D. Lindahl U. Salmivirta M. J. Biol. Chem. 1997; 272: 5518-5524Abstract Full Text Full Text PDF PubMed Scopus (119) Google Scholar, 23Parthasarathy N. Goldberg I.J. Sivaram P. Mulloy B. Flory D.M. Wagner W.D. J. Biol. Chem. 1994; 269: 22391-22396Abstract Full Text PDF PubMed Google Scholar), and antithrombin III requires a 3-O-sulfated N-sulfoglucosamine unit (24Petitou M. Duchaussoy P. Lederman I. Choay J. Sinay P. Carbohydr. Res. 1988; 179: 163-172Crossref PubMed Scopus (53) Google Scholar). Further specificity in the FGF signaling system may also result from either simultaneous or sequential interactions between heparan sulfate and FGF and its receptors (25Allen B.L. Rapraeger A.C. J. Cell Biol. 2003; 163: 637-648Crossref PubMed Scopus (160) Google Scholar, 26Guimond S.E. Turnbull J.E. Curr. Biol. 1999; 9: 1343-1346Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar). Here we have defined specific structural features required for NRG1-heparin interactions and shown the physiological importance of these features in vitro. We demonstrate a clear hierarchy of importance for specific sulfate groups with the N-sulfate group being most important followed by 2-O- and 6-O-sulfate groups for binding to NRG1 and blocking erbB receptor phosphorylation. Consistently, the removal of all of the endogenous sulfate groups with chlorate or selective blockade of N-sulfation with an siRNA directed against the enzyme that promotes N-sulfation resulted in decreased NRG1-induced erbB receptor activation. This specificity provides a means by which tissues can localize and potentiate NRG1 signaling through modifications in GAG composition. Reagents—Recombinant human NRG1 β1-polypeptides were from R&D (Minneapolis, MN). The isolated EGF-like domain corresponded to amino acids 177–246, and the intact IgEGF form of NRG1 corresponded to amino acids 14–246. Sodium chlorate was purchased from Sigma. The heparan sulfate antibody, 10e4, was from Seikagaku. De-N-sulfated, De-2-O-sulfated, and De-6-O-sulfated heparins were prepared as described previously (27Ostrovsky O. Berman B. Gallagher J. Mulloy B. Fernig D.G. Delehedde M. Ron D. J. Biol. Chem. 2002; 277: 2444-2453Abstract Full Text Full Text PDF PubMed Scopus (141) Google Scholar). Different sized heparin fragments from 12 to 2 disaccharides in length were prepared by high resolution gel chromatography of low molecular weight heparin as described previously (28Goger B. Halden Y. Rek A. Mosl R. Pye D. Gallagher J. Kungl A.J. Biochemistry. 2002; 41: 1640-1646Crossref PubMed Scopus (121) Google Scholar). For all of the experiments, the heparins were used at equal weight/volume (μg/ml) concentrations. Cell Cultures and Western Blots for p185 Phosphorylation—L6 cells were used to quantify erbB receptor phosphotyrosine (p185) as described previously (13Li Q. Loeb J.A. J. Biol. Chem. 2001; 276: 38068-38075Abstract Full Text Full Text PDF PubMed Google Scholar). Cells were plated at 50,000 cells/well and grown for 8 days. Chlorate was filter-sterilized and incubated with L6 cells for 2 to the of p185 were by IgEGF NRG1 and of the heparins at for of this was to L6 cells for at in The was and the cells were in of and for as described previously (13Li Q. Loeb J.A. J. Biol. Chem. 2001; 276: 38068-38075Abstract Full Text Full Text PDF PubMed Google Scholar). of this was on a The of the erbB receptors (p185) were by Western using the phosphotyrosine and and by with a of against and was by the of using system on as described previously R.M. Loeb J.A. J. Neurosci. 2004; PubMed Scopus Google Scholar). or NRG1 heparins at were incubated with binding for and on a gel for at as described previously Zhang L. T. B. D.L. A. J. Biol. Chem. 2003; Full Text Full Text PDF PubMed Scopus Google Scholar, Zhang L. D.L. B. J. 2002; PubMed Scopus Google Scholar). the of the to the of NRG1 not the mobility of NRG1 with or the resolution was at and was used for was to a for at NRG1 were by Western using a against the IgEGF The were with a to and to treatment with and the were as siRNA were with from using the The of was of a to L6 cells were incubated with either 2 of siRNA or siRNA in 12 of to of L6 The cells were incubated at for and either with NRG1 for followed by phosphotyrosine Western as were used to reduced by and sulfation activity by with an was isolated using the from to the the was with for The of from the was using the and of from each was using The were using against and 2 and GAG and and and of these a of the of We used with the for followed by of for for and for For we used the system on an and the in gene to for each as described previously R.M. Loeb J.A. J. Neurosci. 2004; PubMed Scopus Google Scholar). was on L6 cells plated at 50,000 cells/well on an and with for at with cells were for with in and incubated with the at a in the block at in and with at for 2 at using a with a of was using in a as described previously (17Li Q. Esper R.M. Loeb J. Mol. Cell. Neurosci. 2004; 26: 558-569Crossref PubMed Scopus (15) Google Scholar). The of of siRNA and cells from were the of cells. cells were reduced to and of the from which an of a to the The S.E. one for NRG1 prepared a of heparins either all of the sulfate groups the N-sulfate groups the 2-O-sulfate groups or the 6-O-sulfate groups to importance in interactions using a that described by and Zhang L. T. B. D.L. A. J. Biol. Chem. 2003; Full Text Full Text PDF PubMed Scopus Google Scholar). oligosaccharides were were and but had a high of structural to the endogenous cell-surface heparan K. N. Res. 2002; 22: PubMed Scopus Google Scholar). This a which complexes on to and The were that NRG1 was to a in mobility in the with NRG1-heparin complexes and a in the mobility of NRG1 by Western that of heparin shift NRG1 in a that is in and NRG1 in a of NRG1 was incubated with or heparin for and on a the heparin was NRG1 in a to form NRG1-heparin complexes as by Western using a NRG1 and in this gel mobility shift we heparin to and The heparin was to shift NRG1 mobility This is not the heparin have low mobility with the which still in the also not shift a critical importance of N-sulfation for NRG1 We examined the importance of 2-O- and on NRG1 binding using heparins each of these Both of these heparins were in not as as and more the This suggests an but equal importance of these two sulfate groups for NRG1 binding The of NRG1-induced in L6 of gel shift using Western for NRG1 as has been for other heparin-binding proteins Zhang L. T. B. D.L. A. J. Biol. Chem. 2003; Full Text Full Text PDF PubMed Scopus Google Scholar, Zhang L. D.L. B. J. 2002; PubMed Scopus Google Scholar), was that NRG1 the at a This in the of the NRG1 NRG1 at the of the Because of this and as a means to the importance of specific heparin sulfation on we examined the ability of these heparins to block erbB receptor phosphorylation in L6 muscle cells. the mechanism by which soluble heparins block erbB receptor activation is not it is a assay to the of NRG1-heparin interactions in a system and has been used previously to the of oligosaccharides J.A. Fischbach G.D. J. Cell Biol. 1995; 130: 127-135Crossref PubMed Scopus (114) Google Scholar). that NRG1-induced erbB receptor phosphorylation is by heparin in a a reduced for each of the heparins to block NRG1-induced erbB receptor phosphorylation. to the gel mobility shift and heparins had at the 2-O- and heparins with The results were in and were as a of low of the heparin had a effect on receptor phosphorylation. This raises the that low of a low heparin can potentiate erbB receptor activation by NRG1 from endogenous sites on the cell on NRG1 and erbB results thus far suggest an importance of the specific sulfation pattern to NRG1-heparin interactions and demonstrate an hierarchy of importance for the sulfate We the importance of length on the NRG1-heparin interactions using a of heparin fragments in from 12 to 2 disaccharides on both and receptor phosphorylation the heparin length was equal weight/volume of these oligosaccharides an shift in the NRG1-heparin The of NRG1 also decreased with heparin that heparins bind NRG1 more we had used of each we have an between the and The of multiple for a length raises the that either the multiple NRG1 bind a heparin or multiple bind a as small as two disaccharides in length were able to shift small of NRG1 on the of the gel on not we the ability of sized heparin fragments to block NRG1-induced receptor phosphorylation to the of NRG1 binding in we as heparin length its ability to NRG1-induced erbB receptor phosphorylation was the length was reduced to disaccharides or was no by Heparan NRG1-induced erbB have shown previously that endogenous HSPGs significantly erbB receptor activation by heparin-binding of NRG1 but not by of NRG1 that the heparin-binding domain (13Li Q. Loeb J.A. J. Biol. Chem. 2001; 276: 38068-38075Abstract Full Text Full Text PDF PubMed Google Scholar). we examined the of endogenous sulfation by L6 muscle cells for with which the of the sulfate for F. K. E. Lindahl U. Salmivirta M. J. Biol. Chem. 1999; Full Text Full Text PDF PubMed Scopus Google Scholar, R. J. Biol. Chem. 2004; Full Text Full Text PDF PubMed Scopus Google Scholar, S. Biochemistry. 2004; PubMed Scopus Google Scholar). the heparin-binding form of NRG1 chlorate treatment reduced receptor phosphorylation by no was with the form that lacks the heparin-binding domain and This the importance of endogenous sulfate groups to potentiate interactions with on the specificity in the and erbB phosphorylation we examined the importance of the endogenous N-sulfate group for interactions in muscle cells. We this by endogenous N-sulfation with an siRNA against N-acetylglucosamine which is the most expressed of the isoforms and a with (4Turnbull J. Drummond K. Huang Z. Kinnunen T. Ford-Perriss M. Murphy M. Guimond S. Biochem. Soc. Trans. 2003; 31: 343-348Crossref PubMed Scopus (23) Google Scholar, K. J. M. K. E. Kjellen L. Biophys. 2002; PubMed Scopus Google Scholar). The treatment of L6 cells with this siRNA resulted in a in using 2, and 3, The in a in the of the N-sulfate groups using that have been shown to require N-sulfate groups for G. B. H. J. Cell Biol. PubMed Scopus Google Scholar). with the siRNA, siRNA a in NRG1-induced receptor phosphorylation in NRG1-induced erbB phosphorylation was with siRNA treatment using the isolated domain of NRG1 and These results suggest an importance of endogenous N-sulfation on interactions that results with the that virtually all of the cells HSPGs and that the of heparin-binding factors to are needed to which factors bind to and concentrate along a cell surface in to requirements during development and in the of HSPGs and specific sulfation are that this be a growth factor is the cell the of that the GAG chains specificity in has been with sulfate each of which a of the muscle during development A. R. J. Neurosci. 2000; PubMed Google Scholar). The localized of two in the and chick and a more of the suggest that this diversity is not but during development through the of K. H. Habuchi H. N. H. Kimata K. J. Biol. Chem. 2004; Full Text Full Text PDF PubMed Scopus Google Scholar). modifications of heparan sulfate chains have also been described that may the sulfation of HSPGs through the activation of specific This the which the of heparan sulfate and promotes signaling D.M. Science. 2001; PubMed Scopus Google Scholar), and the that has been shown to be required to growth factor signaling in cancer J. J. J. R. J. Biol. Chem. 2003; Full Text Full Text PDF PubMed Scopus Google Scholar). Here we have used a of heparins to the specific sulfation requirements for optimal NRG1 two and erbB receptor we a hierarchy of importance with the N-sulfate groups most important for NRG1-heparin interactions followed by the 2-O- and 6-O-sulfate The with the and the De-6-O-sulfated heparins suggest that each to NRG1 binding to a extent. to the sulfation the length was important for optimal We disaccharides in length were for binding to heparin fragments as small as two disaccharides still bind but to a extent. This raises as to the and potential functions of NRG1-heparin interactions with heparins of Because the erbB receptors as both homodimers and binding multiple NRG1 to a may potentiate erbB receptor signaling. The specific sulfation requirements for NRG1-heparin interactions described may also the localization and potentiation of biological functions of NRG1 that with endogenous HSPGs at distinct tissue during development (11Loeb J.A. J. Neurocytol. 2003; 32: 649-664Crossref PubMed Scopus (28) Google Scholar). We that blocking all of the endogenous sulfation with chlorate significantly reduced the ability of NRG1 to erbB receptors in L6 muscle cells. The effect of chlorate treatment on NRG1 activity in muscle cells was not as as that previously all of the with (13Li Q. Loeb J.A. J. Biol. Chem. 2001; 276: 38068-38075Abstract Full Text Full Text PDF PubMed Google or by treatment that and the J. L. M.A. PubMed Scopus Google Scholar). This is not the by chlorate is not and can be selective for specific sulfate groups F. K. E. Lindahl U. Salmivirta M. J. Biol. Chem. 1999; Full Text Full Text PDF PubMed Scopus Google Scholar). The physiological importance of N-sulfation in NRG1-heparin interactions was by endogenous N-sulfation in muscle cells using an siRNA against These cells a in NRG1-induced receptor phosphorylation to the with chlorate treatment that was not with an isolated NRG1 domain that lacks the heparin-binding domain. The of the siRNA treatment was by a in and a in with an that requires an N-sulfated G. B. H. J. Cell Biol. PubMed Scopus Google Scholar, B. H. G. J. Histochem. 1994; PubMed Scopus Google Scholar). The with the siRNA treatment may have resulted from activity or from by other (4Turnbull J. Drummond K. Huang Z. Kinnunen T. Ford-Perriss M. Murphy M. Guimond S. Biochem. Soc. Trans. 2003; 31: 343-348Crossref PubMed Scopus (23) Google Scholar). the siRNA used against with for of and resulted in an in as not potential in these results from the that N-sulfation of of HSPGs is an early in a sequential (4Turnbull J. Drummond K. Huang Z. Kinnunen T. Ford-Perriss M. Murphy M. Guimond S. Biochem. Soc. Trans. 2003; 31: 343-348Crossref PubMed Scopus (23) Google Scholar). by we have also reduced 2-O- and the chlorate result the importance of endogenous the siRNA results suggest that specific sulfation modulate sulfate interactions important for NRG1 localization and activation. Further of with in genes that heparan sulfation M. Irie F. Plump A.S. Tessier-Lavigne M. Yamaguchi Y. Science. 2003; 302: 1044-1046Crossref PubMed Scopus (355) Google Scholar, M. J. K. T. F. I. Kjellen L. E. J. Biol. Chem. 2000; Full Text Full Text PDF PubMed Scopus Google Scholar, G. L. L. Wang B. G. 2000; PubMed Scopus Google be important in these interactions in