JBC Focus on PI3-Kinase with Echelon

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


     

Originally published In Press as doi:10.1074/jbc.M601248200 on July 7, 2006

J. Biol. Chem., Vol. 281, Issue 37, 27178-27189, September 15, 2006
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
281/37/27178    most recent
M601248200v1
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 Email this article to a friend
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
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Duchesne, L.
Right arrow Articles by Fernig, D. G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Duchesne, L.
Right arrow Articles by Fernig, D. G.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

N-Glycosylation of Fibroblast Growth Factor Receptor 1 Regulates Ligand and Heparan Sulfate Co-receptor Binding*Formula

Laurence Duchesne{ddagger}1, Bérangère Tissot§, Timothy R. Rudd{ddagger}2, Anne Dell§3, and David G. Fernig{ddagger}4

From the {ddagger}School of Biological Sciences, University of Liverpool, Liverpool L69 7ZB, United Kingdom and the §Division of Molecular Biosciences, Imperial College, London SW7 2AZ, United Kingdom

Received for publication, February 8, 2006 , and in revised form, June 9, 2006.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The regulation of cell function by fibroblast growth factors (FGF) occurs through a dual receptor system consisting of a receptor-tyrosine kinase, FGFR and the glycosaminoglycan heparan sulfate (HS). Mutations of some potential N-glycosylation sites in human fgfr lead to phenotypes characteristic of receptor overactivation. To establish how N-glycosylation may affect FGFR function, soluble- and membrane-bound recombinant receptors corresponding to the extracellular ligand binding domain of FGFR1-IIIc were produced in Chinese Hamster Ovary cells. Both forms of FGFR1-IIIc were observed to be heavily N-glycosylated and migrated on SDS-PAGE as a series of multiple bands between 50 and 75 kDa, whereas the deglycosylated receptors migrated at 32 kDa, corresponding to the expected molecular weight of the polypeptides. Optical biosensor and quartz crystal microbalance-dissipation binding assays show that the removal of the N-glycans from FGFR1-IIIc caused an increase in the binding of the receptor to FGF-2 and to heparin-derived oligosaccharides, a proxy for cellular HS. This effect is mediated by N-glycosylation reducing the association rate constant of the receptor for FGF-2 and heparin oligosaccharides. N-Glycans were analyzed by mass spectrometry, which demonstrates a predominance of bi- and tri-antennary core-fucosylated complex type structures carrying one, two, and/or three sialic acids. Modeling of such glycan structures on the receptor protein suggests that at least some may be strategically positioned to interfere with interactions of the receptor with FGF ligand and/or the HS co-receptor. Thus, the N-glycans of the receptor represent an additional pathway for the regulation of the activity of FGFs.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The fibroblast growth factors (FGFs)5 and their receptors co-evolved with metazoans (1, 2), a reflection of their central role in mediating cell-cell communication. FGFs regulate the specification of stem cell fate, organogenesis, skeletal growth (3), tissue repair (4) and phosphate metabolism. FGFs are involved in numerous pathologies such as genetic skeletal defects in children (5) and carcinomas in adults (6, 7).

In humans and rodents there are 22 genes encoding more than 30 different FGF proteins and the fgfr 1-4 genes encode over 48 major isoforms of the cognate receptor tyrosine kinase (FGFR) (1, 2, 8). FGFs possess a glycosaminoglycan co-receptor, usually heparan sulfate (HS). The interaction of FGFs with HS is required for the stimulation of cell proliferation, which is initiated by the formation of a complex of FGF ligand, HS co-receptor and FGFR (9-11). Not surprisingly for a family of key regulatory ligands, the activation of FGF signaling is subjected to multiple regulatory inputs, some of which operate at the level of the assembly of the receptor-ligand complex, whereas others operate inside the cell, on the active receptor complex. Regulatory inputs that operate extracellularly include partial selectivity in the choice of ligand by certain receptor isoforms. FGF-7 is the most specific FGF ligand, because it only interacts with the IIIb isoform of FGFR2; other FGFs show varying degrees of promiscuity in terms of their FGFR preference (12). The HS chain mediates a second extracellular regulatory input. Specific sequences in HS can allow only a restricted subset of FGF-FGFR interactions to lead to cell proliferation (13-16) and this is likely to occur in vivo (17, 18). A third extracellular regulatory input is likely to be the assembly of the FGF receptor ligand system into different complexes that have different signaling potentials. This is suggested by the identification of two different crystal structures of complexes of FGFR·FGF·heparin oligosaccharide (19, 20) and by biophysical (21) and biochemical evidence (9, 22). Finally, genetic evidence indicates that mutation of an asparagine residue in the extracellular domain of FGFR2 and FGFR3 increases its activity, to result in skeletal growth defects, which has been linked to the disruption of N-glycosylation (23, 24). The latter observations suggest that N-glycosylation of FGFR may constitute a fourth regulatory input.


Figure 1
View larger version (33K):
[in this window]
[in a new window]
  		FIGURE 1.
Recombinant FGFR1-ST and FGFR1-GPI proteins: schematic structure and oligonucleotides used to produce expression vectors. A, schematic representation of FGFR1-ST and FGFR1-GPI proteins. Two recombinant forms of the extracellular ligand binding domain of rat FGFR1-IIIc (amino acids 120-368 comprising the acid box and immunoglobulin (Ig) loops D2 and D3) were produced. A soluble form of the receptor possessed a poly-histidine tail (6xhis) sequence followed by a thrombin cleavage site at the N terminus and a Factor Xa cleavage site followed by a Strep-Tag II sequence (ST) at the C terminus and is termed FGFR1-ST. A membrane-bound form of the receptor possessed the same N-terminal His6 sequence and thrombin cleavage site, but after the Factor Xa cleavage site a GPI anchor attachment sequence (GPI) replaced the Strep-Tag II sequence. This form of the receptor is termed FGFR1-GPI. B, oligonucleotides used to obtain pcDNA-ht-fgfr1-xgpi and pcDNA-ht-fgfr1-xst cDNA plasmids. Using as a template the pcDNA-h-fgfr1-gpi, two-step PCR was performed using ThroF, ThroR, pHindF, and pXbaR primers to introduce a thrombin cleavage site sequence. Then, the FXaF, FXaR, pHindF, and pXbaR primers were used to introduce the Factor Xa cleavage site sequence. The resulting pcDNA-ht-fgfr1-xgpi plasmid was used as a template with pHindF and pSTR primers to obtain pcDNA-ht-fgfr1-xst.

 
To ascertain whether N-glycans on FGFR may regulate aspects of the assembly of the FGF ligand-receptor complex, the extracellular ligand binding domain of FGFR1-IIIc was expressed in Chinese Hamster Ovary (CHO) cells. We observed that it was heavily N-glycosylated and that multiple glycoforms of the receptor protein existed. The N-glycans were found to inhibit the interaction between FGFR1 and its FGF-2 ligand and HS co-receptor. The results suggest that N-glycosylation of FGFR1 represents an important additional pathway for the regulation of activity of the FGF receptor.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Materials—Restriction enzymes and Vent polymerase were from New England Biolabs (Wilbury Way Hitchin, Herts, UK). Lipofectamine and Ni-NTA-agarose resin were from Invitrogen. CHO cells were the generous gift of Dr. D. Moss, University of Liverpool and were cultured in Dulbecco's modified Eagle's medium/Nutrient Mixture F-12 Ham (DMEM/F-12 Ham) supplemented with 5% (v/v) fetal calf serum (Invitrogen). Recombinant human FGF-2 was produced as described (25). Heparin-derived dodecasaccharides were prepared from partial heparinase I digests of pig mucosal heparin and were obtained from Iduron (Manchester, UK). Bis[sulfosuccinimidyl] suberate, BS3, was from Perbio (Chester, UK). N-(3-Dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride, EDC, was from Fluka (Buchs, Switzerland). Thrombin, streptavidin, 11-mercaptoundecanoic acid, ethanolamine, hydrazine monohydrate and antibody to His6 were from Sigma. Strep-Tactinhorseradish peroxidase and Strep-Tactin-Sepharose resin were from IBA (Goettingen, Germany). Factor Xa was purchased from Roche Diagnostics (Lewes, East Sussex, UK) and N-glycanase (peptide N-glycosidase F) from Prozyme (Ely, UK).

Plasmid Constructions—Two recombinant forms of the extracellular ligand binding domain of rat FGFR1-IIIc (26) (amino acids 120-368 comprising the acid box and immunoglobulin loops D2 and D3) were produced. A soluble form of the receptor possessed a poly-histidine tail (His6) sequence followed by a thrombin cleavage site at the N terminus and a Factor Xa cleavage site followed by a Strep-Tag II sequence (27) at the C terminus and is termed FGFR1-ST (Fig. 1A). A membrane-bound form of the receptor possessed the same N-terminal His6 sequence and thrombin cleavage site, but after the Factor Xa cleavage site a GPI anchor attachment sequence replaced the Strep-Tag II sequence. This form of the receptor is termed FGFR1-GPI (Fig. 1A). Constructs encoding these two proteins were obtained using as a template pcDNA-h-fgfr1-gpi (gift from Dr. M. Seve, University of Liverpool) which corresponds to the rat fgfr1-IIIc sequence (amino acids 120-368) flanked by secretion signal and polyhistidine (His6) sequences in 5' and a sequence encoding a GPI attachment sequence in 3' (28). By performing two-step polymerase chain reactions (PCR) (29) using sets of appropriate primers, sequences encoding cleavage sites for thrombin and for Factor Xa, as well as Strep-Tag II were introduced (Fig. 1B).

Expression and Purification of FGFR1-ST and FGFR1-GPI Proteins—pcDNA-ht-fgfr1-xgpi and pcDNA-ht-fgfr1-xst constructs were linearized by digestion with BglII and transfected into CHO cells using Lipofectamine. Transfectants were selected by resistance to G418. For the soluble FGFR1-ST protein, clones expressing around 1 µg of receptor protein/ml culture medium were identified after plating into 96-well plates and Western blotting of culture medium with anti-His6. For the membrane-bound FGFR1-GPI, FACS using a Coulter Epics Altra flow cytometer (Beckman Coulter, Fullerton CA) was used to isolate a population of cells expressing the highest levels of receptor protein by labeling the cells with anti-His6 antibody and a second FITC-conjugated antibody. The population of cells identified and sorted by FACS were then plated into 96-well plates and clones expressing the highest levels of FGFR1-GPI were identified by Western blotting of cell extracts. The highest levels of expression of receptor protein for all clones was obtained by growing the cells in Dulbecco's modified Eagle's medium/Nutrient Mixture F-12 Ham supplemented with 0.5% (v/v) fetal calf serum for 72 h. In some experiments cells were cultured in the presence of 3 µg/ml tunicamycin for 24-48 h and recombinant proteins analyzed by SDS-PAGE followed by Western blot.

For FGFR1-ST protein expression, cells were cultured in Dulbecco's modified Eagle's medium/Nutrient Mixture F-12 Ham supplemented with 0.5% (v/v) fetal calf serum in a rotor cell 12 prototype (Powell Brothers Ltd., Ormskirk, Lancs, UK). Medium was collected every 5 days. After flow dialysis of the culture medium, FGFR1-ST was purified by nickel chelation and Strep-Tactin-Sepharose chromatography. Vivaspin filtration unit (cut-off 30 kDa, Vivascience) was then used to concentrate and exchange the purified FGFR1-ST into PBS (phosphate-buffered saline, 140 mM NaCl, 10 mM Na2HPO4/NaH2PO4, pH 7.2). Concentration of receptor protein was determined by measurement of the absorbance at 280 nm.

Removal of N-glycans of purified protein was performed by N-glycanase digestion in PBS buffer following the manufacturer's recommendations. To ensure equal amounts of protein were used in all experiments, a single aliquot of receptor protein was divided into two, one was treated with N-glycanase and the other mock-treated.

Optical Biosensor Binding Assays—FGF-2 was immobilized on aminosilane surfaces using bissulfosuccinimidyl suberate (BS3 as the cross linker following the manufacturer's recommendations (NeoSensors, Sedgefield, UK). No more than 300 arc s FGF-2 was immobilized on the surface (1 arc s = 1/3600°, 600 arc s = 1 ng protein/mm2). Heparin-derived dodecasaccharides were biotinylated at their reducing ends and immobilized on streptavidin surfaces, as described (22, 30).

Binding assays were carried out in PBS supplemented with 0.02% (v/v) Tween 20 (PBST) at 20 °C following previously described methods (13, 14, 22, 30-33) with minor modifications. A single binding assay consisted of adding the soluble binding partner or ligate in 1-3 µl PBST to a cuvette containing 22-24 µl or 37-39 µl PBST. The association reaction was followed until binding was at least 90% of the calculated equilibrium value, usually between 150 s and 230 s. The cuvette was then washed three times with 50 µl of PBST to initiate the dissociation of bound FGFR1-ST. The FGF-2 and oligosaccharide-derivatized surfaces were regenerated by washing twice with 50 µl of 2 M NaCl in 10 mM sodium phosphate buffer (10 mM Na2HPO4/NaH2P04, pH 7.2), PBST and 20 mM HCl, which removed 98-100% of bound FGFR1-ST. Binding parameters were calculated using the non-linear curve fitting program FASTFit (NeoSensors). Each binding assay yielded four binding parameters, which are the slope of initial rate of association, the on-rate constant (kon) and the extent of binding, all calculated from the association phase, and the off-rate constant (koff, equivalent to the dissociation rate constant, kd), calculated from the dissociation phase (13, 14, 22, 30-33). All biosensor experiments were carried out at least two times on at least two different surfaces.

Data Analysis—The determination of binding kinetics in optical biosensors can be prone to artifactual second phase binding sites, due either to rates of diffusion of soluble ligate approaching or exceeding the rate of association, or to steric hindrance at the binding surface (33). Binding assays were designed to avoid such artifacts (13, 14, 22, 30-33). Thus, limiting amounts of ligand were immobilized on the sensor surface, whereas the slope of initial rate, kon and the extent of binding were only determined at low concentrations of ligate, and koff was measured at higher concentrations of ligate, to avoid steric hindrance and rebinding artifacts. A single site model fitted the data at least as well as more complex models, was covered by at least 90% of the data, and, therefore, was used to calculate all binding parameters. The equilibrium dissociation constant (Kd) was calculated both from the ratio of the kd and ka and from the extent of binding, to provide an estimate of the self-consistency of the results.

Quartz Crystal Microbalance-Dissipation (QCM-D)—A self-assembled monolayer was formed on a QCM-D (Q-Sense AB, Västra Frölunda, Sweden) gold crystal by overnight incubation in 11-mercaptoundecanoic acid (0.1 M in ethanol). The surface was then washed with ethanol, water (MilliQ) and then finally with PBS. All subsequent reactions were carried out in PBS. A hydrazide surface was formed by activating the terminal carboxylic acids with a wash of 300 µl of 0.25 M EDC followed by incubation in 300 µl of 0.2 M hydrazine monohydrate and 0.25 M EDC for 1 h. The surface was then washed vigorously with PBS.

To immobilize a heparin-derived tetradecasaccharide, the hydrazide surface was incubated overnight in 300 µlof5 µg/ml heparin-derived tetradecasaccharide, a kind gift of Dr A. K. Powell, University of Liverpool. FGF-2 was immobilized on the hydrazide surface by cross-linking with BS3: 3 injections of 300 µl, 0.1 M BS3 over 10 min, followed by two injections of 300 µl, 37 µg/ml FGF-2. After the latter injection, the surface was incubated for 50 min before any unreacted BS3 was blocked with ethanolamine, 0.1 M, 300 µl, for 10 min.

In a binding experiment, 300 µl of the deglycosylated or glycosylated FGFR1-ST (48.5 nM) was introduced into the cell and left to incubate for 10 min at 20 °C. The surface was then washed with PBS. The amount of FGFR1-ST bound was determined by comparing the frequency response before and after the addition of the receptor. This frequency change was converted into a mass change by using the Sauerbrey equation (34), with a conversion co-efficient of 17.7 ng cm-2 Hz-1. The surface was regenerated by washing with 0.5 M NaCl and 0.1 M HCl.

Pull-down Experiments—In heparin pull-down experiments, 2 µg of glycosylated and deglycosylated FGFR1-ST in 0.5 ml of PBS was incubated overnight at 4 °C with 5 µl of heparin-agarose (Bio-Rad). Beads were collected by centrifugation at 13,000 x g for 3 min and washed twice with PBS by centrifugation. Protein in the supernatant was concentrated by freeze-drying. Supernatants and final pellets were adjusted in volume and proteins were analyzed by SDS-PAGE followed by silver nitrate staining.

Analysis of the N-Glycans on FGFR1-ST—FGFR1-ST protein (50 µg) was reduced for 1 h at 37°C in 50 mM Tris-HCl buffer (pH 8.5) containing a 4-fold excess of dithiothreitol and carboxymethylated with a 2-fold molar excess of iodoacetic acid for 1 h at room temperature in the dark. Following dialysis at 4 °C for 72 h against 4 x 4.5 liters of cold 50 mM ammonium bicarbonate, pH 7.5, and lyophilization, FGFR1-ST was digested with trypsin (Thr1246, Sigma) at a 12:1 ratio (w/w) in 50 mM ammonium bicarbonate (pH 8.5) for 18 h at 37 °C. The reaction was stopped by adding a few drops of acetic acid to the solution. The sample was lyophilized prior to its dissolution in 150 µl (5% (v/v)) of acetic acid and purified using a SepPak cartridge C18 (Waters Corp), as described (35). The purified glycopeptides were then digested with PNGase-F (Roche Applied Science, 1365177) in 50 mM ammonium bicarbonate (pH 8.5) containing 4.5 units of enzyme at 37 °C over 18 h. The sample was lyophilized, and the released N-glycans were purified using a SepPak cartridge C18 (Waters Corp) (35). Permethylation and sample clean-up were performed using the sodium hydroxide protocol, as described previously (35).

MALDI-TOF data were acquired using a Perseptive Biosystems Voyager DE-STRTM mass spectrometer in the reflector mode with delayed extraction. MS/MS data were acquired using a 4800 MALDI TOF/TOF (Applied Biosystems) mass spectrometer. The collision energy was set to 1 kV, and air was used as collision gas. Samples were dissolved in 10 µl of methanol and mixed at a 1:1 ratio (v/v) with 2,5-dihydrobenzoic acid as matrix.

In Silico Glycosylation of the FGF Receptor—In silico glycosylation of the FGF receptor using GlyProt (36) employed two different models of the structures of immunoglobin loops D2 and D3 of FGFRs as input: the symmetric "two end" model of the FGFR1-IIIc·FGF-2·heparin hexasaccharide crystal structure (PDB ID: 1FQ9) and the asymmetric model of FGFR2-IIIc·FGF-1·heparin decasaccharide crystal structure (PDB ID: 1E0O). N-glycans were incorporated that correspond to structures identified by MALDI-TOF mass spectrometry. For a given model, the same glycan structure was used for all six potential N-glycosylation sites of the FGFR1-ST. PyMOL viewer software was used to visualized the model three-dimensional structures (DeLano, W. L. (2002) The PyMOL Molecular Graphics System).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
N-Glycosylation of Recombinant FGFR1 Proteins—Culture medium from CHO cells transfected with FGFR1-ST contained polypeptides immunoreactive to anti-His6 that migrated at an apparent molecular size of 50-70 kDa. When the same cells were cultured in the presence of tunicamycin (3 µg/ml), an inhibitor of N-glycosylation, a single immunoreactive band of 32 kDa was observed (Fig. 2A). Because the calculated molecular mass of FGFR1-ST is 31.5 kDa, this result suggested that the FGFR1-ST secreted by the transfected cells was heavily N-glycosylated. To determine if N-glycosylation was intrinsic to the protein or the consequence of it being secreted rather than membrane-bound, CHO cells were also transfected with FGFR1-GPI. In this case cell-associated immunoreactivity migrated as a series of bands between 45 and 75 kDa. When cultured in the presence of tunicamycin, cells expressed an immunoreactive product that migrated as a single band of 32 kDa, the expected size of FGFR1-GPI (Fig. 2A). FGFR1-ST was purified from CHO cell culture medium by nickel chelation and Strep-Tactin-Sepharose chromatography. After a subsequent separation by SDS-PAGE a series of bands between 50 and 75 kDa were revealed using silver staining. Preincubation of this material with N-glycanase, an enzyme which removes N-glycans, resulted in a single silver-stained band of 32 kDa (Fig. 2B). Thus, the multiple bands observed after separation of the FGFR1-ST reflect the existence of multiple glycoforms of the protein, which is pure as judged by silver staining.


Figure 2
View larger version (18K):
[in this window]
[in a new window]
  		FIGURE 2.
Expression of FGFR1-ST and FGFR1-GPI proteins in CHO cells. A, CHO cells expressing soluble FGFR1-ST or cell membrane-associated FGFR1-GPI were cultured in the presence (+) or absence (-) of 3 µg/ml tunicamycin. The medium (FGFR1-ST) and cell lysate (FGFR1-GPI) were analyzed by SDS-PAGE followed by Western blot using anti-His6. When cells are cultured in tunicamycin, FGFR1-ST and FGFR1-GPI migrate at the size expected for the non-glycosylated proteins. B, purified FGFR1-ST protein was treated with and without N-glycanase and analyzed by SDS-PAGE followed by silver nitrate staining. After N-glycanase treatment, FGFR1-ST migrates at the size expected for the non-glycosylated protein.

 
In cells treated with tunicamycin, an inhibitor of N-glycosylation, FGFR1-ST, and FGFR1-GPI still traversed the secretory pathway, because immunoreactive deglycosylated FGFR1-ST and FGFR1-GPI were detected in the culture medium and cells, respectively (Fig. 2A). The level of immunoreactive FGFR1-ST secreted into the culture medium is reduced in the presence of tunicamycin. This may reflect, at least in part, the inhibition of cell proliferation by tunicamycin and the fact that equivalent amounts of culture medium rather than cells are compared. When equal amounts of cellular protein are compared, as in the case with FGFR1-GPI, tunicamycin is observed to reduce the level of immunoreactive FGFR1-GPI to a far lesser extent. Given that the relative affinities of the antibody to His6 for glycosylated and deglycosylated receptor protein are not known, these results demonstrate that N-glycosylation is not essential for the folding and secretion of FGFR1-ST and FGFR1-GPI. However, N-glycosylation may contribute to the efficiency of these processes.

Binding of FGFR1-ST to FGF-2 and Heparin Oligosaccharides—In a first experiment the level of binding of FGFR1-ST to immobilized FGF-2 was measured using an optical biosensor. The relatively low level of binding observed was caused by the loss of much of the binding activity of the FGF-2 after immobilization (results not shown). The results show that more deglycosylated FGFR1-ST bound to FGF-2 than N-glycosylated FGFR1-ST (Fig. 3A). Similarly, more deglycosylated FGFR1-ST than N-glycosylated FGFR1-ST bound to a biotinylated heparin dodecasaccharide immobilized on the sensor surface (Fig. 3B). Optical biosensors measure changes in refractive index. Thus, it was possible that the differences observed may be caused by the different refractive indices of the N-glycosylated and the deglycosylated receptor. As a control for this possibility, a binding experiment was performed in a QCM-D (37), in which changes in the frequency of oscillation of a quartz crystal are caused by changes in its mass because of the adsorption/desorption of molecules to the crystal surface (34). Deglycosylated FGFR1-ST bound to a greater extent to immobilized FGF-2 (Fig. 3C) and heparin tetradecasaccharide (Fig. 3D). Therefore, these differences in levels of binding must reflect an effect of the N-glycans on the molecular interactions rather than on the refractive index of the receptor protein.


Figure 3
View larger version (17K):
[in this window]
[in a new window]
  		FIGURE 3.
Interaction of glycosylated and deglycosylated FGFR1-ST with immobilized FGF-2 and heparin-derived oligosaccharides and with heparin in solution. Experiments with immobilized ligands were performed in an optical biosensor (A and B) and in a QCM-D (C and D). FGFR1-ST was deglycosylated by treatment with N-glycanase ("Experimental Procedures"). A, FGF-2 was immobilized on an aminosilane surface and the binding of identical amounts of glycosylated and deglycosylated FGFR1-ST (230 nM) was measured. B, dodecasaccharide biotinylated at the reducing end was captured on a streptavidin surface and the binding of the same amount of glycosylated and deglycosylated FGFR1-ST (290 nM,) was measured. For both surfaces, association data were acquired until binding was at least 90% of the maximal level and analyzed with one-site binding model to determine the extent of binding model ("Experimental Procedures"). Results are mean ± S.E. of triplicates (A, glycosylated FGFR1-ST; B) or duplicate (A, deglycosylated FGFR1-ST) of the extent of binding calculated from fitting data to a one site model. Nonspecific binding was not detectable. C, FGF-2 was immobilized on a QCM-D crystal surface and the binding of glycosylated and deglycosylated FGFR1-ST (49 nM) was measured. D, a heparin-derived tetradecasaccharide was immobilized on a QCM-D crystal surface and the binding of glycosylated and deglycosylated FGFR1-ST (49 nM) was measured. QCM-D results are the means ± S.E. of triplicates. E, heparin-agarose was incubated overnight with glycosylated or deglycosylated FGFR1-ST and then collected by centrifugation. P is the pellet, and SN the supernatant, which were analyzed by SDS-PAGE followed by Western blot with antibody to His6.

 


Figure 4
View larger version (40K):
[in this window]
[in a new window]
  		FIGURE 4.
Binding of deglycosylated FGFR1-ST to FGF-2. The binding kinetics of deglycosylated FGFR1-ST to immobilized FGF-2 were measured as described under "Experimental Procedures." Data shown are the result of one representative experiment of three. A, deglycosylated FGFR1-ST receptor at different concentrations was added to a FGF-derivatized cuvette containing 25 µl of PBST, and the association reaction was followed for 250 s, by which time at least 90% of the fitted one site binding curve was covered. The concentration of deglycosylated FGFR1-ST in nM is indicated. Data were collected three times a second. B-F, the distribution of the data points (jagged line) around a one site binding model (horizontal line at 0 arc s) is shown for each of the concentrations of FGFR1-ST used in the binding assay in panel A. Instrument noise is ± 0.5 arc s. Concentrations of deglycosylated FGFR1-ST are 57.5 nM (B), 115 nM (C), 230 (D), 460 nM (E), 690 nM (F). G, linear relationship between the slope of initial rate of association and concentration of deglycosylated FGFR1-ST. H, linear relationship between kon, determined from a one-site model, and concentration of deglycosylated FGFR1-ST.

 
Kinetics of Interaction of FGFR1-ST to FGF-2 and Heparin Oligosaccharides—To characterize the binding of the glycosylated and deglycosylated receptor in more detail, the kinetics of their interactions with FGF-2 (Figs. 4 and 5) and a heparin-derived dodecasaccharide (supplemental data, Figs. S1 and S2) were determined. As the concentration of FGFR1-ST increased, so did the observed level of binding (Figs. 4A and 5A, supplemental data, Figs. S1A and S2A). Analysis of the binding curves with a one site model resulted in a random distribution of the data about the model (Figs. 4, B-F and 5, B-F; supplemental data, Figs. S1, B-F and S2, B-F). A plot of the slope of initial rate against FGFR1-ST concentration was fitted by a straight line (Figs. 4G and 5G, supplemental data, Figs. S1G and S2G), as was a plot of the observed on-rate kon against the concentration of FGFR1-ST (Figs. 4H and 5H; supplemental data, Figs. S1H and S2H). In addition, the data covered at least 90% of the curve described by the single site model. Taken together, these analyses indicate that, at the concentrations used for the experiment, the interaction of deglycosylated and glycosylated FGFR1-ST with FGF-2 and heparin dodecasaccharides was monophasic. The N-glycosylated FGFR1-ST possessed a relatively slow ka for immobilized FGF-2 and an affinity, calculated from the kinetic parameters of the interaction, of 2.9 µM (Table 1). In contrast, the ka for the deglycosylated FGFR1-ST was almost 5-fold faster. The higher affinity of the deglycosylated receptor for FGF-2 was, therefore, largely because of a faster ka (Table 1). The affinity of these interactions calculated from the maximum extent of binding Kd (equilibrium) was similar in each case to the value calculated from the kinetic parameters, indicating that these binding data are internally self-consistent.


View this table:
[in this window]
[in a new window]
  		TABLE 1
Kinetics of glycosylated and deglycosylated FGFR1-ST binding to immobilized FGF-2 and a heparin-derived oligosaccharide Binding parameters were determined in an optical biosensor ("Experimental Procedures").

 


Figure 5
View larger version (39K):
[in this window]
[in a new window]
  		FIGURE 5.
Binding of glycosylated FGFR1-ST to FGF-2. The binding kinetics of glycosylated FGFR1-ST to immobilized FGF-2 were measured as described under "Experimental Procedures." Data shown are the result of one representative experiment of three. A, glycosylated FGFR1-ST receptor at different concentrations was added to a FGF-derivatized cuvette containing 25 µl of PBST, and the association reaction was followed for 250 s, by which time at least 90% of the fitted one site binding curve was covered. The concentration of glycosylated FGFR1-ST in nanomolar is indicated. Data were collected three times a second. B-F, the distribution of the data points (jagged line) around a one site binding model (horizontal line at 0 arc s) is shown for each of the concentrations of FGFR1-ST used in the binding assay in panel A. Instrument noise is ± 0.5 arc s. Concentrations of glycosylated FGFR1-ST are 115 nM (B), 230 nM (C), 460 (D), 760 nM (E), 1500 nM (F). G, linear relationship between the slope of initial rate of association and concentration of glycosylated FGFR1-ST. H, linear relationship between kon, determined from a one-site model, and concentration of glycosylated FGFR1-ST.

 
The kinetic parameters for the interaction of glycosylated and deglycosylated FGFR1-ST to a heparin dodecasaccharide indicate that the N-glycosylated FGFR1-ST possesses a relatively slow ka for immobilized DP 12 and an affinity of 0.33 µM. These values are in the same range as those obtained for the interaction with heparin of FGFR1-IIIc, expressed as an Fc fusion protein (31). In contrast, the ka of the deglycosylated FGFR1-ST for the dodecasaccharide was almost 2-fold faster. The higher affinity of the deglycosylated receptor for heparin was, therefore, again largely because of a faster ka (Table 1). The affinity of these interactions calculated from the maximum extent of binding Kd (equilibrium) was, for the glycosylated FGFR1-ST, different to the value calculated from the kinetic parameters. Nevertheless, both the Kd (kinetic) and the Kd (equilibrium) indicate a lower affinity of the glycosylated receptor for heparin compared with that of the deglycosylated receptor.

Pull-down of FGFR1-ST Glycoforms by Heparin—The differences in affinities between the N-glycosylated and deglycosylated FGFR1-ST observed in the optical biosensor binding experiments (Table 1) are largely because of differences in ka, which is a measure of the probability of a molecular collision resulting in a binding reaction. The differences in ka could, therefore, reflect the fact that the productive collision profile for N-glycosylated FGFR1-ST is narrower than for the deglycosylated receptor, perhaps because of steric hindrance by the N-glycans. Alternatively, the most heavily glycosylated species of FGFR1-ST might not bind FGF-2 or heparin at all. In this case, the reduced ka observed with the N-glycosylated receptor would merely reflect the substantially lower concentration of the minor glycoforms that have substantially less N-glycans (Fig. 2A) and which would then be the sole species capable of binding FGF-2 and the heparin dodecasaccharide. To distinguish between these possibilities, with respect to heparin binding, "affinity pull-down" experiments were designed. When heparin-agarose beads were used to pull-down FGFR1-ST, differences between glycosylated and deglycosylated FGFR1-ST were observed. Indeed, a substantial proportion of the glycosylated FGFR1-ST was not pulled-down by heparin, whereas all deglycosylated receptor bound heparin (Fig. 3E). However, within the limits of detection, all glycosylated species of FGFR1 bound heparin. Therefore, this result suggests that in the case of the interaction between FGFR1-ST and heparin, N-glycans on the receptor strongly reduce the probability of binding, but they do not prevent it.

Structures of the N-Glycans on FGFR1-ST—The N-glycans of FGFR1-ST protein were released using PNGase F, permethylated, and subjected to MALDI-TOF mass profiling, complemented by MS/MS sequencing of selected ions using MALDI-TOF-TOF instrumentation. In the MALDI-TOF profile obtained in the low m/z range (Fig. 6A) and in the high m/z range (Fig. 6B), the majority of molecular ions correspond to complex type N-glycans, with compositions varying from 0 to 4 NeuAc residues, 0-1 Fuc residues, 5-8 Hex residues and 4-7 HexNAc residues. These structures (Fig. 6) were deduced from the compositional and MS/MS data, taking into account the glycosylation capabilities of CHO cells, in particular their inability to fucosylate the antennae of complex N-glycans. Thus, the major peak at m/z 2605 corresponds to a bi-antennary complex type structure carrying one sialic acid and a core fucose (composition: NeuAc1Fuc1Hex5HexNAc4). Its disialylated counterpart is present at m/z 2969 (Fig. 6A). Analogous, fully sialylated, tri- and tetra-antennary glycans give the signals at m/z 3777 and 4587, respectively (Fig. 6B).

MS/MS analysis showed that a significant portion of the glycans have extended antennae. Thus, the MS/MS spectrum of m/z 3055 exhibited fragment ions corresponding to both biantennary and tri-antennary structures (see annotations on Fig. 6A), including ions at m/z 2144, corresponding to loss of a terminal HexHexNAcHexHexNAc, and at m/z 2130 corresponding to loss of two terminal HexHexNAc moieties, respectively (data not shown). Further evidence for both bi- and tri-antennary structures was provided by peaks at m/z 1738 (loss of a SiaHexHexNAcHexHexNAc moiety) and m/z 1769, which corresponds to the concomitant loss of a terminal SiaHexHexNac and a terminal HexHexNAc. Similarly, the abundant molecular ion at m/z 3416 was shown by MS/MS to be a mixture of bi- and tri-antennary glycans (see annotations on Fig. 6B). The diagnostic fragment ions demonstrating the presence of the triantennary isomeric form (m/z 2953 corresponding to loss of a terminal HexHexNAc and m/z 2129 corresponding to the concomitant loss of a terminal SiaHexHexNAc and a terminal HexHexNAc) are more abundant than those diagnostic of the extended bi-antennary glycan (m/z 2143 corresponding to loss of a terminal SiaHexHexNAcHexHexNAc), suggesting that the tri-antennary structure dominates. Analysis of the minor m/z 3504 ion suggested the presence of the bi-, tri-, and tetra-antennary structures shown in Fig. 6B. For example, evidence for extended antennae was provided by the presence of an intense peak at m/z 2593 arising from loss of a terminal HexHexNAcHexHexNAc moiety. The structures shown on m/z 3867, 4226, and 4314 (Fig. 6B) were similarly assigned.

N-Glycans of the FGFRs: Sequence and Structure—The extracellular ligand binding domain of FGFR1-IIIc possesses 6 potential N-glycosylation sites. These glycosylation sites are relatively well conserved among the FGFR family (Fig. 7, A and B). Indeed, they are all present in FGFR1 and in FGFR2 and those on the D3-loop (n3, 4, 5, 6) are common to all receptors (Fig. 7, A and B). Several or all of these sites are likely to be glycosylated, because there is a substantial change in migration of glycosylated FGFR1-ST observed upon SDS-PAGE (Fig. 2). Models of the crystal structures of FGFR1-IIIc·FGF-2·heparin hexasaccharide (PDB ID: 1FQ9; Fig. 7C) and FGFR2-IIIc·FGF-1·heparin decasaccharide (PDB ID: 1E0O; Fig. 7D) show that if all 6 N-glycosylation sites were occupied, the receptor protein would be heavily cloaked in polysaccharide.

Considering the symmetric model (Fig. 7C), three asparagine residues are noteworthy with respect to the negative regulation of binding of FGFR1-ST to FGF-2 and to heparin by N-glycans. The FGF ligand binding site identified by site-directed mutagenesis and co-crystallography is adjacent to Asn227 in D2 immunoglobulin loop and Asn317 in D3 immunoglobulin loop (19, 20, 38, 39). N-Glycosylation of these residues would be likely to modify the ligand binding properties of the receptor. The heparin binding site has been located to the D2 immunoglobulin loop (19, 20) where Asn227 and Asn240 are situated. Although they are not immediately adjacent to this site, their sheer size may be sufficient to inhibit heparin binding. Moreover, if these N-glycans were terminated by sialic acid, its negative charge could compete with the binding of the anionic heparin. Asn264 is positioned strategically, such that it would prevent dimerization of FGFR1. Considering the asymmetric model (Fig. 7D), Asn228 and Asn265 are adjacent to the FGF binding site and would be likely to inhibit ligand binding. In this model, no potential N-glycosylation seems to be able to interfere with the binding of heparin to the receptor or with receptor dimerization. It must be noted that N-glycans are flexible structures, and the torsion angle of their bond to the receptor asparagines is unknown. Therefore, it is likely that at least some of the N-glycans modeled here may be oriented differently. Moreover, these polysaccharide structures are highly mobile and may dynamically occupy a far larger space than identified in these models and consequently the occupation of fewer N-glycosylation sites may achieve a similar cloaking effect on the protein surface.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
The importance of protein glycosylation is becoming widely recognized through studies on the folding, trafficking, and functions of proteins. The two recombinant forms of the FGFR1-IIIc, corresponding to the extracellular ligand binding domain the receptor, are heavily N-glycosylated when produced in CHO cells. This glycosylation occurs whether the receptor is membrane-bound (FGFR1-GPI) or secreted into the culture medium (FGFR1-ST) (Fig. 2). Although the N-glycosylated asparagines on FGFR1 have yet to be fully identified, SDS-PAGE and mass spectrometry experiments show a large number of glycosylated isoforms (Figs. 2 and 6). The potential N-glycosylation sites in immunoglobulin loops D2 and D3 of FGFRs are strongly conserved among the family, and data from studies on FGFR4 show that, even in Sf9 insect cells, two of these sites are N-glycosylated (equivalent to Asn264, Asn296 in FGFR1, Fig. 7B) (40). Moreover, a study on FGFR3 strongly suggests that, at least Asn328 is glycosylated in human blood lymphocytes (23). Taken together, this suggests that at least three of the potential N-glycosylation sites in FGFR1 may be glycosylated. This is supported by the observation that the mobility of glycosylated FGFR1-ST on SDS-PAGE is substantially altered compared with that of the deglycosylated protein (Fig. 2).


Figure 6
View larger version (28K):
[in this window]
[in a new window]
  		FIGURE 6.
MALDI-TOF mass spectrum of FGFR1-ST N-glycans. Glycans were released by PNGase F, permethylated, and subjected to Sep-Pak cleanup ("Experimental Procedures"). Data from one of the fractions collected upon SepPak cleanup are shown. A, low mass range N-glycans (from 1500 to 3250 Da). B, high mass range N-glycans (from 3250 to 5000 Da). Scheme assignments are based on the precise fit between composition calculations and the m/z (z = 1) ratio of the molecular ions detected. These schemes represent the most likely structures taking into account the biosynthetic pathways and the enzyme repertoire of CHO cells. Ions at 3055, 3416, 3504, 3867, 4226, and 4314 m/z were subjected to MALDI TOF/TOF MS/MS analysis to discriminate structures with extended antennae from those with additional short antennae. The unlabeled satellite peaks 30 mass units above each major signal correspond to artifacts from the permethylation reaction (56). The three N-glycan structures used for the in silico glycosylation of the FGF receptor (Fig. 6) are boxed.

 


Figure 7
View larger version (52K):
[in this window]
[in a new window]
  		FIGURE 7.
N-Glycans of the FGFRs: sequence and structure. A, sequence alignment of Human FGFR1-4 proteins corresponding to the FGFR1-ST recombinant receptor. Amino acids are in single letter code. All FGFRs are numbered starting with the methionine encoded by the initiation codon. The six potential N-glycosylation sites (NX(S/T) where X is any amino acid except P, D, or E), identified using NetNglyc server, are colored in orange. *, known mutations of the consensus N-glycosylation sites abolishing the potential for N-glycosylation. FGFR2 S267P and FGFR2 N331I lead to Crouzon syndrome and FGFR3 N328I leads to hypochondroplasia (23, 24). B, positions of the asparagines potentially glycosylated in loops D2-D3 for FGFR1-4. C and D, two different models of the crystal structures of D2-D3 loop FGFRs. C, symmetric "two end" model of the FGFR1-IIIc·FGF-2·heparin hexasaccharide crystal structure (PDB ID: 1FQ9). D, the asymmetric model of FGFR2-IIIc·FGF-1·heparin decasaccharide crystal structure (PDB ID: 1E0O). FGFR monomers are colored in blue, the FGF ligands in green, and the heparin oligosaccharides in red. The asparagines (N, single letter code) potentially glycosylated are in gray (C1, D1). Using these crystal structures, models were generated incorporating some of the N-glycan structures identified by MALDI-TOF mass spectrometry (Fig. 6). C2 and D2, symmetric "two end" and "asymmetric" model glycosylated with a bi-antennary N-glycan structure with a bisecting fucose and decorated with a terminal sialic acid (E; Fig. 6A) at their 6 potential N-glycosylation sites. Additional models of N-glycosylated FGFR are available in the supplemental data (supplemental data, Fig. S3).

 
The extracellular, two Ig loop form of FGFR1-IIIc produced here (Fig. 1A) is still secreted in the presence of tunicamycin (Fig. 2A). This suggests that the N-glycans on the receptor are not essential for protein folding and trafficking through the secretory apparatus, although these sugars may contribute to the efficiency of these processes. However, the deglycosylated receptor possesses a higher affinity for FGF-2 and for a heparin dodecasaccharide than the glycosylated one (Figs. 3, 4, 5; supplemental data, Figs. S1 and S2). This effect is exerted at the level of ka (Table 1). The affinity pull-down experiment (Fig. 3E) suggests that, within the limits of detection, all glycosylated species bound heparin. Therefore, the differences in ka are likely to reflect the fact that the productive collision profile for N-glycosylated FGFR1-ST is narrower than for the deglycosylated receptor.

Analysis of the receptor N-glycans by MALDI-TOF mass spectrometry and by MS/MS sequencing of selected ions identified at least 29 bi-, tri-, and tetra-antennary oligosaccharides decorated with terminal sialic acid and core fucose (Fig. 6). N-glycosylation in CHO cells is similar, but not identical to that in humans. For example, CHO cells lack the fucosyltransferase responsible for terminal fucosylation, whereas in humans the genes encoding the enzymes responsible for transfer of terminal N-glycolyl neuraminic acid and {alpha}-galactose contain frame shifts that prevent synthesis of active enzyme. The consequent differences in N-glycosylation are considered to be minor, which is evidenced by the fact that CHO cells are the system of choice for the production of therapeutic glycoproteins (41, 42). Moreover, it is important to note that there can be far greater differences in N-glycosylation patterns observed for a single glycoprotein produced by a human cell under different stimulatory and physiological conditions (41, 43, 44). Modeling of bi-antennary, tri-antennary and tetra-antennary N-glycans on the potential N-glycosylation sites of the FGFR1 demonstrates that, if they are all occupied, the receptor protein is heavily cloaked in polysaccharide (Fig. 7, C and D and supplemental data, Fig. S3). Therefore, one mechanism whereby N-glycans can reduce the probability of binding is by steric hindrance. Indeed, some of the N-glycans are positioned such that they could interfere with the binding of FGF (Asn227 and Asn317), the binding of heparin (Asn227 and Asn240) or the dimerization of the receptor (Asn264) (Fig. 7, C and D).

Asn296 and Asn330 are distant from the molecular interfaces predicted by both the symmetric and asymmetric model structures and so do not at first sight seem to be of importance for the assembly of the complex of FGFR·FGF·HS. Nevertheless, both these N-glycosylation sites are conserved in all FGFRs, suggesting that they may be of functional importance. Moreover, a naturally occurring mutation at the equivalent position of N330 in FGFR2 and in FGFR3 causes skeletal growth defects through overactivation of the receptor and it has been suggested that the phenotype associated with this mutation is the direct consequence of the disruption of the N-glycosylation site (23, 24). Studies on the regulation of FGF signaling by HS and on FGF ligand-receptor complex assembly have suggested that the FGFR may possess other partners, which also bind HS (13). Recently, it has been shown that neuropilin-1, a protein involved in angiogenesis and axon guidance, is able to bind the FGFR1 and the FGF ligand (32), as well as HS (45). Other partners of the FGFR have been identified: anosmin-1, cadherins, and N-CAM (46-50). Finally, there is evidence that the FGFR·FGF·HS complex itself might assemble into an entity larger than a receptor dimer (21). Therefore, although existing models derived from crystal structures do not support a direct involvement of glycans attached to Asn296 and Asn330 in the regulation of the formation of the simplest FGF receptor-ligand assemblies, it seems likely that the presence of N-glycans at these position may interfere with the formation of higher order molecular complexes of the receptor or with the interaction of the FGFR with partners others than the FGF ligand and the glycosaminoglycan co-receptor.

The present results demonstrate that N-glycosylation of the FGFR influences the assembly of the FGF·HS·FGFR complex. A number of lines of evidence suggest that heterogeneity of N-glycosylation, which may arise during receptor biosynthesis or at the cell surface, may produce subclasses of FGFR with different activities. Firstly, cross-linking of 125I-FGF (FGF-1 or FGF-2), used to identify cell surface receptors, identified two species of ~120 and 140 kDa (51-53), which were subsequently interpreted to be the three (140 kDa) and two (120 kDa) Ig loop isoforms of FGFR. However, given the existence of multiple glycoforms of FGFR that cause a ~20 kDa spread of the receptor protein on SDS-PAGE, e.g. Fig. 2A, if the 125I-FGF bound all glycoforms equally, these two isoforms might be expected to appear as a smear covering 30-40 kDa. Secondly, tyrosine phosphorylated FGFR has generally been identified in cells transfected to overexpress the receptor. In this case, immunoprecipitated FGFR probed on a Western blot with antibodies to FGFR gives a broad smear analogous to that Fig. 2, whereas the same blot probed with antibodies to phosphorylated tyrosine produces a much tighter band of immunoreactivity (54). Thirdly, the constitutive ligand-independent activation of FGFR2 caused by a C342Y mutation is accompanied by a reduction in receptor N-glycosylation (55). These observations, along with those on the Asn330 mutant in FGFR2 and FGFR3, which causes overactivation of the receptor (23, 24), are given a mechanistic underpinning by the present work: N-glycosylation reduces at least the association rate constants of two key binary interactions: FGFR·FGF and FGFR·heparin. A reduction in receptor glycosylation would, therefore, be expected to lead to an increased activity of the receptor at limiting concentrations of FGF and/or activating HS co-receptor.

The binding of FGF ligand causes the dimerization of the FGFR in the membrane, which activates the intracellular kinase of the receptor. This key event is subject to regulation by multiple molecular determinants: FGF ligand, isoform of the FGFR, structure of the HS co-receptor and other accessory molecules (32, 46, 48-50). The present results indicate that N-glycosylation of the FGFR may also regulate the assembly of active receptor complexes and so FGF signaling events, providing an entirely novel and additional layer of control over this key ligand receptor system. Because N-glycosylation is dynamic and regulated at the level of individual glycoproteins (43), changes in N-glycosylation of FGFR operating at the level of biosynthesis would be expected to have a substantial effect on the activity of the receptor. There is also a distinct possibility that selective deglycosylation of FGFR may occur at the cell surface, which would enhance the activity of the receptor. Such regulation of receptor activity by N-glycosylation would not be identified in the molecular analyses commonly employed in the study of embryonic development and of model tumors and human cancers, although global changes in N-glycosylation have a long standing association with the progression of malignant disease. Therefore, it is possible that hitherto undetected changes in FGFR N-glycosylation may have profound effects on the activity of the receptor.


   FOOTNOTES
 
* This work was supported in part by grants from the Biotechnology and Biological Sciences Research Council (BBSRC), the Cancer and Polio Research Fund (to D. G. F.), the European Union Marie Curie Programme (to L. D.), the Human Frontiers Science Programme (to D. G. F.), the Wellcome Trust (to A. D.), and by the RCUK Basic Technology Programme (UK Glyco-chips Consortium grant, to B. T.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

Formula The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1-S3. Back

2 Recipient of a BBSRC studentship. Back

3 A BBSRC professorial fellow. Back

1 To whom correspondence may be addressed: School of Biological Sciences, Biosciences Bldg., Crown St., University of Liverpool, Liverpool L69 7ZB, UK. Tel.: 44-151-795-4471; Fax: 44-151-795-4406; E-mail: lduchesn{at}liv.ac.uk. 4 A North West Cancer Research Fund Professor of Biological Chemistry. To whom correspondence may be addressed: School of Biological Sciences, Biosciences Bldg., Crown St., University of Liverpool, Liverpool L69 7ZB, UK. Tel.: 44-151-795-4471; Fax: 44-151-795-4406; E-mail: dgfernig{at}liv.ac.uk.

5 The abbreviations used are: FGF, fibroblast growth factor; FGFR1, FGF receptor-1; FGFR1-ST, extracellular domain of FGFR1-IIIc with a C-terminal Strep-Tag II; GPI, glycosylphosphatidylinositol; FGFR1-GPI, extracellular domain of FGFR1-IIIc with a C-terminal GPI anchor; MALDI-TOF, matrix-assisted laser desorption /ionization-time of flight; PBS, phosphate-buffered saline; QCM-D, quartz crystal microbalance-dissipation; PDB, Protein Data Bank; FACS, fluorescent-activated cell sorting; CHO, Chinese hamster ovary; HS, heparan sulfate. Back


   ACKNOWLEDGMENTS
 
We thank Dr. David Spiller of the Liverpool Centre for Cell Imaging for FACS analysis.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Ornitz, D. M., and Itoh, N. (2001) Genome Biol. 2, 1-12[Medline] [Order article via Infotrieve]
  2. Itoh, N., and Ornitz, D. M. (2004) Trends Genet. 20, 563-569[CrossRef][Medline] [Order article via Infotrieve]
  3. Ornitz, D. M. (2005) Cytokine Growth Factor Rev. 16, 205-213[CrossRef][Medline] [Order article via Infotrieve]
  4. Detillieux, K. A., Sheikh, F., Kardami, E., and Cattini, P. A. (2003) Cardiovasc. Res. 57, 8-19[Abstract/Free Full Text]
  5. Ornitz, D. M., and Marie, P. J. (2002) Genes Dev. 16, 1446-1465[Free Full Text]
  6. Freeman, K. W., Gangula, R. D., Welm, B. E., Ozen, M., Foster, B. A., Rosen, J. M., Ittmann, M., Greenberg, N. M., and Spencer, D. M. (2003) Cancer Res. 63, 6237-6243[Abstract/Free Full Text]
  7. van Rhijn, B. W., Lurkin, I., Radvanyi, F., Kirkels, W. J., van der Kwast, T. H., and Zwarthoff, E. C. (2001) Cancer Res. 61, 1265-1268[Abstract/Free Full Text]
  8. Groth, C., and Lardelli, M. (2002) Int. J. Dev. Biol. 46, 393-400[Medline] [Order article via Infotrieve]
  9. Delehedde, M., Seve, M., Sergeant, N., Wartelle, I., Lyon, M., Rudland, P. S., and Fernig, D. G. (2000) J. Biol. Chem. 275, 33905-33910[Abstract/Free Full Text]
  10. Delehedde, M., Lyon, M., Sergeant, N., Rahmoune, H., and Fernig, D. G. (2001) J. Mammary Gland Biol. Neoplasia 6, 253-273[CrossRef][Medline] [Order article via Infotrieve]
  11. Eswarakumar, V. P., Lax, I., and Schlessinger, J. (2005) Cytokine Growth Factor Rev. 16, 139-149[CrossRef][Medline] [Order article via Infotrieve]
  12. Ornitz, D. M., Xu, J., Colvin, J. S., McEwen, D. G., MacArthur, C. A., Coulier, F., Gao, G., and Goldfarb, M. (1996) J. Biol. Chem. 271, 15292-15297[Abstract/Free Full Text]
  13. Ostrovsky, O., Berman, B., Gallagher, J., Mulloy, B., Fernig, D. G., Delehedde, M., and Ron, D. (2002) J. Biol. Chem. 277, 2444-2453[Abstract/Free Full Text]
  14. Rahmoune, H., Gallagher, J. T., Rudland, P. S., and Fernig, D. G. (1998) J. Biol. Chem. 273, 7303-7310[Abstract/Free Full Text]
  15. Merry, C. L., Bullock, S. L., Swan, D. C., Backen, A. C., Lyon, M., Beddington, R. S., Wilson, V. A., and Gallagher, J. T. (2001) J. Biol. Chem. 276, 35429-35434[Abstract/Free Full Text]
  16. Guimond, S. E., and Turnbull, J. E. (1999) Curr. Biol. 9, 1343-1346[CrossRef][Medline] [Order article via Infotrieve]
  17. Kato, M., Wang, H., Kainulainen, V., Fitzgerald, M. L., Ledbetter, S., Ornitz, D. M., and Bernfield, M. (1998) Nat. Med. 4, 691-697[CrossRef][Medline] [Order article via Infotrieve]
  18. Chang, Z., Meyer, K., Rapraeger, A. C., and Friedl, A. (2000) Faseb J. 14, 137-144[Abstract/Free Full Text]
  19. Pellegrini, L., Burke, D. F., von Delft, F., Mulloy, B., and Blundell, T. L. (2000) Nature 407, 1029-1034[CrossRef][Medline] [Order article via Infotrieve]
  20. Schlessinger, J., Plotnikov, A. N., Ibrahimi, O. A., Eliseenkova, A. V., Yeh, B. K., Yayon, A., Linhardt, R. J., and Mohammadi, M. (2000) Mol. Cell 6, 743-750[CrossRef][Medline] [Order article via Infotrieve]
  21. Harmer, N. J., Ilag, L. L., Mulloy, B., Pellegrini, L., Robinson, C. V., and Blundell, T. L. (2004) J. Mol. Biol. 339, 821-834[CrossRef][Medline] [Order article via Infotrieve]
  22. Delehedde, M., Lyon, M., Gallagher, J. T., Rudland, P. S., and Fernig, D. G. (2002) Biochem. J. 366, 235-244[Medline] [Order article via Infotrieve]
  23. Winterpacht, A., Hilbert, K., Stelzer, C., Schweikardt, T., Decker, H., Segerer, H., Spranger, J., and Zabel, B. (2000) Physiol. Genomics 2, 9-12[Abstract/Free Full Text]
  24. Steinberger, D., Mulliken, J. B., and Muller, U. (1996) Hum. Mutat. 8, 386-390[CrossRef][Medline] [Order article via Infotrieve]
  25. Ke, Y., Wilkinson, M. C., Fernig, D. G., Smith, J. A., Rudland, P. S., and Barraclough, R. (1992) Biochim. Biophys. Acta 1131, 307-310[Medline] [Order article via Infotrieve]
  26. Yan, G. C., Wang, F., Fukabori, Y., Sussman, D., Hou, J. Z., and Mckeehan, W. L. (1992) Biochem. Biophys. Res. Commun. 183, 423-430[CrossRef][Medline] [Order article via Infotrieve]
  27. Schmidt, T. G., Koepke, J., Frank, R., and Skerra, A. (1996) J. Mol. Biol. 255, 753-766[CrossRef][Medline] [Order article via Infotrieve]
  28. Lodge, A. P., Howard, M. R., McNamee, C. J., and Moss, D. J. (2000) Brain. Res. Mol. Brain. Res. 82, 84-94[Medline] [Order article via Infotrieve]
  29. Horton, R. M., and Pease, L. R. (1991) in Directed Mutagenesis, A Practical Approach (McPherson, M. J., ed), IRL Press, Oxford
  30. Delehedde, M., Lyon, M., Vidyasagar, R., McDonnell, T. J., and Fernig, D. G. (2002) J. Biol. Chem. 277, 12456-12462[Abstract/Free Full Text]
  31. Powell, A. K., Fernig, D. G., and Turnbull, J. E. (2002) J. Biol. Chem. 277, 28554-28563[Abstract/Free Full Text]
  32. West, D. C., Rees, C. G., Duchesne, L., Patey, S. J., Terry, C. J., Turnbull, J. E., Delehedde, M., Heegaard, C. W., Allain, F., Vanpouille, C., Ron, D., and Fernig, D. G. (2005) J. Biol. Chem. 280, 13457-13464[Abstract/Free Full Text]
  33. Fernig, D. G. (2001) in Proteoglycan Protocols (Iozzo, R. V., ed), Humana Press, Totowa, NJ
  34. Sauerbrey, G. (1959) Zeitschrift Fur Physik 155, 206-222[CrossRef]
  35. Sutton-Smith, M., and Dell, A. (2005) Cell Biology: A Laboratory Handbook, Academic Press, San Diego
  36. Bohne-Lang, A., and von der Lieth, C. W. (2005) Nucleic Acids Res. 33, 214-219[CrossRef]
  37. Rodahl, M., Hook, F., Fredriksson, C., Keller, C. A., Krozer, A., Brzezinski, P., Voinova, M., and Kasemo, B. (1997) Faraday Discussions 107, 229-246
  38. Plotnikov, A. N., Hubbard, S. R., Schlessinger, J., and Mohammadi, M. (2000) Cell 101, 413-424[CrossRef][Medline] [Order article via Infotrieve]
  39. Springer, B. A., Pantoliano, M. W., Barbera, F. A., Gunyuzlu, P. L., Thompson, L. D., Herblin, W. F., Rosenfeld, S. A., and Book, G. W. (1994) J. Biol. Chem. 269, 26879-26884[Abstract/Free Full Text]
  40. Tuominen, H., Heikinheimo, P., Loo, B. M., Kataja, K., Oker-Blom, C., Uutela, M., Jalkanen, M., and Goldman, A. (2001) Protein Expr. Purif. 21, 275-285[CrossRef][Medline] [Order article via Infotrieve]
  41. Goochee, C. F., Gramer M. J., Andersen, D. C., Bahr, J. C., and Rasmussen, J. R. (1991) Bio/Technology 9, 1347-1355[CrossRef][Medline] [Order article via Infotrieve]
  42. Baker, K. N., Rendall, M. H., Hills, A. E., Hoare, M., Freedman, R. B., and James, D. C. (2001) Biotechnol. Bioeng. 73, 188-202
  43. Gagneux, P., and Varki, A. (1999) Glycobiology 9, 747-755[Abstract/Free Full Text]
  44. Goochee, C. F., and Monica, T. (1990) Biotechnology 8, 421-427[CrossRef][Medline] [Order article via Infotrieve]
  45. Mamluk, R., Gechtman, Z., Kutcher, M. E., Gasiunas, N., Gallagher, J., and Klagsbrun, M. (2002) J. Biol. Chem. 277, 24818-24825[Abstract/Free Full Text]
  46. Gonzalez-Martinez, D., Kim, S. H., Hu, Y., Guimond, S., Schofield, J., Winyard, P., Vannelli, G. B., Turnbull, J., and Bouloux, P. M. (2004) J. Neurosci. 24, 10384-10392[Abstract/Free Full Text]
  47. Bryant, D. M., Wylie, F. G., and Stow, J. L. (2005) Mol. Biol. Cell 16, 14-23[Abstract/Free Full Text]
  48. Williams, E. J., Furness, J., Walsh, F. S., and Doherty, P. (1994) Neuron 13, 583-594[CrossRef]