Identification of the Key Amino Acids of Glial Cell Line-derived Neurotrophic Factor Family Receptor α1 Involved in Its Biological Function*

Glial cell line-derived neurotrophic factor (GDNF) plays a critical role in neurodevelopment and survival of midbrain dopaminergic and spinal motor neurons in vitro and in vivo. The biological actions of GDNF are mediated by a two-receptor complex consisting of a glycosylphosphatidylinositol-linked cell surface molecule, the GDNF family receptor α1 (GFRα1), and receptor protein tyrosine kinase Ret. Although structural analysis of GDNF has been extensively examined, less is known about the structural basis of GFRα1 function. In this study, based on evolutionary trace method and relative solvent accessibility prediction of residues, a set of trace residues that are solvent-accessible was selected for site-directed mutagenesis. A series of GFRα1 mutations was made, and PC12 cell lines stably expressing different GFRα1 mutants were generated. According to the survival and differentiation responses of these stable PC12 cells upon GDNF stimulation and the GDNF-GFRα1-Ret interaction assay, residues 152NN153, Arg259, and 316SNS318 in the GFRα1 central region were found to be critical for GFRα1 binding to GDNF and eliciting downstream signal transduction. The single mutation R259A in the GFRα1 molecule simultaneously lost its binding ability to GDNF and Ret. However N152A/N153A or S316A/N317A/S318A mutation in the GFRα1 molecule still retained the ability to bind with Ret. These findings suggest that distinct structural elements in GFRα1 may be involved in binding to GDNF and Ret.

complexes do not yield x-ray quality co-crystals, and these complexes are frequently beyond the current limits of multidimensional NMR spectroscopy. Thus, structural knowledge is often limited to either receptor or ligand alone, which by itself lacks explicit binding site information. Even when the structure of a complex is available, it is very difficult to deduce the relative contribution of each individual residue to the total binding energy. In fact, the binding site is a distinct subset of the contact sites. Mutagenesis studies have shown that less than half of the residues buried in the binding interface may contribute ϳ90% of the total binding energy (1). Similarly, in the case of the neurotrophin system, three basic residues provide the critical binding determinants for interaction with their p75 receptor (2,3). Mutational analysis, therefore, remains a mainstay of molecular determination of binding interfaces between receptor-ligand complexes.
The glial cell line-derived neurotrophic factor family ligands include GDNF 1 (4), neurturin (5), artemin (6), and persephin (7). They are critical regulators of neurodevelopment (4) and support the survival of midbrain dopaminergic (6,8) and spinal motor neurons (9,10) in vitro and in animal disease models, making them attractive therapeutic candidates for treatment of neurodegenerative diseases (11,12). GDNF is also an inducer and branching factor of ureteric buds during kidney development (13) and plays an important role in sperm cell development (14). The neurotrophic and morphogenic activities of GDNF are mediated by its interaction with a multicomponent receptor complex formed by the Ret receptor tyrosine kinase (15)(16)(17)(18) and a glycosylphosphatidylinositol (GPI)-anchored accessory receptor, GDNF family receptor ␣1 (GFR␣1) (19,20). Four different GFR␣ receptors have been identified (GFR␣1-4), and each of them prefers binding to one of the glial cell line-derived neurotrophic factor family ligands (21). GDNF binds GFR␣1 with preferred high affinity. The unprocessed precursor of rat GFR␣1 contains 468 amino acids, with a signal peptide at the amino terminus and a stretch of 38 hydrophobic amino acids at the carboxyl terminus. The hydrophobic cluster at the carboxyl terminus is preceded by a group of three small amino acids (Ala, Ser, and Ser), defining a cleavage/binding site for GPI linkage, suggesting that GFR␣1 is an extracellular protein that is attached to the outer cell membrane (20). GDNF, GFR␣1, and Ret bind together and subsequently lead to Ret dimerization and activation of the Ret tyrosine kinase. The original model of GDNF signaling proposes a stringent division of labor between GFR␣1 and Ret receptors, in which the latter transduces intracellular signaling but cannot bind ligand on its own, and the former binds ligand but is thought not to transduce signals in the absence of Ret. However, GDNF mutants deficient in GFR␣1 binding are still able to activate Ret normally, suggesting that at least some Ret molecules are weakly associated with GFR␣1 before GDNF binding (22). The precise details of receptor complex formation and stoichiometry of components within the complex remain unclear. The GDNF-GFR␣1-Ret complex provides an attractive system in which to investigate protein-protein interactions involved in the assembly of multisubunit receptor complexes. The x-ray crystal structure of GDNF (23) and its distinct structural elements involved in GFR␣1 binding (12,22,24) have been determined. However, less is known about the structure and function analysis of GFR␣1. One study has reported that the central region of GFR␣ receptors constitutes a novel binding domain for cysteine-knot superfamily ligands. The carboxyl-terminal segments adjacent to the central domain are necessary and have modulatory function in ligand binding (25). In our work, a set of residues in the central and carboxylterminal domain of GFR␣1 were investigated for their involvement in functional receptor complex formation and signal transduction by alanine-scanning mutagenesis. The selected residues are evolutionarily conserved residues identified by the evolutionary trace method (26) and also have relatively high solvent accessibility in PHD prediction (27). After generating stable PC12 cell lines containing this series of GFR␣1 mutants, we examined GDNF binding, Ret activation, and differentiation and survival responses to GDNF. Our results indicate that several amino acids in the GFR␣1 receptor are critical for binding with GDNF and Ret and mediating the neurotrophic function of GDNF.

EXPERIMENTAL PROCEDURES
Reagents-The restriction endonucleases and T4 DNA ligase were purchased from MBI. TRIzol reagent was obtained from Invitrogen. Omniscript reverse transcriptase kits were from Qiagen. Pryobest DNA polymerase was obtained from Takara Biotechnology Co., Ltd. Mini-MTM DNA extraction system and gel extraction miniprep kit were from VIOGENE. Plasmid pBluescript SK(ϩ), pCDNA3-GFR␣1, and pCDNA3-Ret were generated previously in our laboratory. The poly-Llysine, fetal calf serum, horse serum, and Dulbecco's modified Eagle's medium were purchased from Invitrogen. PC12 cells were the generous gift from the Cell Bank of Shanghai Institute of Cell Biology, Chinese Academy of Sciences. FuGENE 6.0 was purchased from Roche Applied Science. Anti-phosphotyrosine monoclonal antibody pY99, anti-GDNF mouse monoclonal antibody B-8, anti-GFR␣1 rabbit polyclonal antibody H-70, and anti-Ret goat polyclonal C-20 antibodies were from Santa Cruz Biotechnology. Fluorescein isothiocyanate-conjugated chicken anti-rabbit IgG and rhodamine-conjugated donkey anti-goat IgG were from Chemicon International, Inc. GDNFs were prepared as described previously (28). All other chemicals were purchased from BBI.
Evolutionary Trace Analysis, Secondary Structure Prediction, and Site-directed Mutagenesis-Based on the principle of molecular evolution, sequence and structure information can be integrated by evolutionary analysis to determine functional sites in the protein. This method is now called evolutionary trace (ET) (26). The improved ET method (29,30) was used to identify the functional epitopes in the members of the GFR␣ family. A total of 37 GFR␣ primary sequences was compiled following a BLAST search with the sequence of rat GFR␣1 over the NCBI non-redundant data base. Multiple sequence alignment and dendrogram construction were then carried out by the ClustalX program (31). Secondary structure and residues relative to solvent accessibility predictions of rat GFR␣1 were deduced using the PHD predict program (27).
The full-length of rat GFR␣1 cDNA was subcloned into pCDNA3.0. Single-stranded DNA from this plasmid was used as a template for oligonucleotide-based site-directed mutagenesis as described previously (32). All mutations were confirmed by DNA sequence analysis.
Generation of Stable PC12 Cells Containing Mutant GFR␣1 Recep-tors-PC12 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum, 5% horse serum, and 0.35% glucose. All cells were grown at 37°C in 5% CO 2 . 1.3 ϫ 10 5 PC12 cells were plated in each well of a 12-well plate (Corning Glass) that had been coated with poly-L-lysine prior to transfection. After overnight incubation, the monolayer cell density would achieve 50 -80% confluence. pCDNA3.0-GFR␣1 wild type or GFR␣1 mutants were co-transfected with pCDNA3.0-Ret plasmid using FuGENE 6.0, and stable PC12 cell lines were selected in medium containing G418. Differentiation and Survival Assay-For the differentiation assay, 2 ϫ 10 4 PC12 cells stably expressing GFR␣1 and Ret constructs were added to each well of a 12-well plate pre-coated with poly-L-lysine. After attachment, the cells were exposed to 100 ng/ml GDNF. Seven days later, the effects of GDNF on cell differentiation were determined. PC12 cells possessing one or more neurites of a length more than twice the diameter of the cell body were scored as positive. Each value was the mean Ϯ S.E. sampled from three independent experiments. For the survival assay, PC12 cells were seeded as described above. After attachment, the cells were switched to serum-free medium plus 100 ng/ml GDNF. Cell survival was quantified by staining with 10 g/ml fluorescein diacetate (Sigma) after 48 h as described previously (33). Viable cells were determined by fluorescence microscopy. The value of the viable cell number in culture with the medium-containing serum represented our corrected 100% initial survival. Results are expressed as the percentage of cell counts with respect to the 100% initial value and show the mean Ϯ S.E. of the percentages from three independent experiments.
Reverse Transcription-PCR Analysis-Cells were lysed in TRIzol, and the RNA concentration was measured photometrically. First strand cDNA was synthesized from 1 g of total RNA using (dT) 16 -18 . The PCR primer pair 5Ј-cacaaggccctcaggcagttcttcgac-3Ј and 5Ј-ttgaattgcatttttgagacaagtatt-3Ј corresponding to the specific part of GFR␣1 was amplified to a 453-bp product. After an initial denaturing step of 5 min at 95°C, the amplification was carried out for 30 cycles at 95°C for 30 s, 57°C for 30 s, and 72°C for 30 s. A final amplification for 10 min at 72°C finished the PCR. The product was cloned into the pBSK vector and verified by DNA sequencing.
Immunofluorescence Cell Staining-Cells grown on coverslips were washed briefly with PBS and fixed with 4% paraformaldehyde for 20 min. Cells were blocked with 10% normal goat serum at room temperature for 20 min and then incubated with primary antibodies at 4°C overnight followed by fluorescent secondary antibodies at room temperature for 45 min. Coverslips were mounted with 90% glycerol in PBS and examined by fluorescent microscopy.
Circular Dichroism and Unfolding Measurements-Recombinant proteins of wild type GFR␣1 and its mutants GFR␣1(N152A/N153A), GFR␣1(Q247A/D248A/S249A), GFR␣1(R259A), and GFR␣1(S316A/ N317A/S318A) were expressed in Escherichia coli and purified as described previously (36). Far-UV CD measurements were performed on a JASCO-715 spectropolarimeter with a scan speed of 20 nm/min and a response time of 0.125 s. The path length of the cuvette was 1 mm. Each spectrum was the average of at least three scans. The protein concentration was 0.2 mg/ml (in 1 ϫ PBS buffer, pH 7.6), except for the samples in the unfolding experiments. The ellipticity at 220 nm under different concentrations of GdmCl was recorded for the unfolding transition profile.

Site-directed Mutagenesis of Selected Residues in GFR␣1-
The primary sequences of GFR␣ family members were retrieved by BLAST search and used for the evolutionary trace analysis. The dendrogram of these sequences separated them into four main clusters that each correlated with a different subgroup in this family from GFR␣1 to ␣4 (data not shown). The globally invariant residues that are conserved in all the GFR␣ family members were identified; these residues are likely to contribute to important structural domains or some shared functions for the whole family. The class-specific residues of GFR␣1, which are conserved in GFR␣1 subgroup but are variable in other GFR␣ subgroups, are considered to be important for delineating the functional specificities of GFR␣1 subgroup. The GFR␣ family receptors have a unique pattern of cysteine residues and lack many of the domains most commonly present in other receptors. Thus, this receptor family likely represents a structurally novel receptor class. Secondary structure and relative solvent accessibility predictions of residues of rat GFR␣1 were done by PHD prediction. Our secondary structure predictions results were similar to those reported previously (25). According to the domain boundaries defined by those authors, we also divided the GFR␣1 molecule into aminoterminal (three predicted helices), central (four predicted helices and two ␤-strands), and carboxyl-terminal (two predicted helices) domains of roughly 100, 200, and 100 residues, respectively (Fig. 1). The residues targeted for site-directed mutagenesis are trace residues, both globally invariant residues and class-specific residues, in the central domain or carboxyl-terminal segments adjacent to the central domain of GFR␣1 with relatively high solvent accessibility. The selected residues that fit the criteria for mutagenesis are shown in Fig. 1. These residues may be on the protein surface and specify the functional epitopes in GFR␣1 that interact with other molecules. Among them, Asn 152 , Arg 171 , Arg 197 , Arg 259 , and Ser 318 are globally invariant residues in the whole family, and Asn 153 , Lys 206 , Gln 247 , Asp 248 , Ser 249 , Glu 270 , Asp 284 , Ser 316 , Asn 317 , Leu 326 , Lys 327 , Asp 334 , and Lys 339 are class-specific residues of the GFR␣1 subgroup. These residues were mutated into alanine, either individually or in combinations of two or three residues (Table I). Alanine is best suited to the scanning approach because it can accommodate most elements of the secondary structure of proteins and results in minimal structural distortion.
The Effects of GDNF on PC12 Cells Stably Expressing GFR␣1 and Ret-By stable selection, a series of PC12 cell lines stably expressing GFR␣1 mutants and Ret were generated. In order to investigate the key amino acids of GFR␣1 involved in its biological function, the effects of GDNF on cell differentiation and survival were evaluated by treating these PC12 cells with 100 ng/ml GDNF. Seven days after treatment, compared  (Fig. 2, C and  D)). However, GDNF could not induce neurite outgrowth in PC12-GFR␣1(N152A/N153A)-Ret, PC12-GFR␣1(R259A)-Ret, and PC12-GFR␣1(S316A/N317A/S318A)-Ret stable cells (Fig.  2, B and D). The survival-promoting effects of GDNF on PC12stable cells were consistent with differentiation responses of PC12 stable cells to GDNF (Fig. 3). Two days after treatment, The predicted secondary structure of rat GFR␣1 is displayed with the GPI anchor at the carboxyl terminus. Predicted ␣-helices are represented as cylinders and ␤-strands as arrows. The amino-terminal domain, the central domain, and the carboxyl-terminal domains are colored dark gray, light gray, and gray, respectively. The selected residues targeted by sitedirected mutagenesis are indicated.

TABLE I
Summary of mutants effects of GFR␣1 on GDNF and Ret binding, mediating Ret phosphorylation and neurotrophic function of GDNF GFR␣1 or GFR␣1 mutants and RET gene were co-transfected into PC12 cells to construct a series of PC12 stable cell lines. After GDNF stimulation, cell lysates were immunoprecipitated with anti-GFR␣1 antibodies and Western blot with anti-GDNF and anti-Ret antibodies, respectively, to detect binding with GDNF and Ret and immunoprecipitated with anti-Ret antibodies and Western blot with anti-phosphotyrosine antibodies to detect Ret phosphorylation. Neurite outgrowth cells and survival cells were calculated, respectively, as described under "Experimental Procedures" to detect which residue was critical in mediating neurotrophic function of GDNF. ϩϩ indicates normal effects (like wild type GFR␣1); ϩ indicates reduced effects; Ϫ indicates no effects. GFR␣1  Ret stable cells in serum-free medium (Fig. 3).

Determination of Expression of GFR␣1 or GFR␣1 Mutants and Ret in PC12
Cells-To eliminate the possibility that the absence of response of PC12-GFR␣1(N152A/N153A)-Ret, PC12-GFR␣1(R259A)-Ret, and PC12-GFR␣1(S316A/N317A/S318A)-Ret stable cells to GDNF may be due to lack of expression of these GFR␣1 mutants, RT-PCR and Western blot analyses were performed to confirm appropriate expression in these PC12 cell lines. The transcription of GFR␣1 mutants was ascertained by RT-PCR, utilizing GFR␣1-specific primers that produced an expected amplified product of ϳ450 bp. The results showed GFR␣1 mutants mRNA were all transcribed in these PC12 stable cell lines (Fig. 4). In addition, Western blot analysis of PC12 cell lysates was performed using anti-GFR␣1 and anti-Ret antibodies to confirm GFR␣1 and Ret protein expression in these stable cell lines. The Western blots confirmed that all GFR␣1 mutants were expressed in the stable PC12 cell lines, and the expression levels between GFR␣1 and the various mutants were equivalent (Fig. 5). Ret was also expressed in all stably transfected PC12 cell lines (Fig. 6).
Immunofluorescence Cell Staining-It is well known that GFR␣1 protein is linked to the plasma membrane by a glycosylphosphatidylinositol anchor. Immunofluorescence microscopy was used to investigate whether GFR␣1 mutants localized to the cell surface. The results showed that GFR␣1 staining was bright and sharp along the cell surface of the stably transfected PC12 cells, with no staining in the control group (Fig. 7). These results suggested that the GFR␣1 mutant proteins expressed in transfected PC12 cells were properly localized on the cell surface.
Phosphorylation and GDNF-GFR␣1-Ret Interaction-To determine which residues of GFR␣1 are involved in binding to GDNF or Ret and to influence the activation of Ret tyrosine kinase, a series of stable PC12 cells were treated with GDNF for 15 min, followed by immunoprecipitation with anti-GFR␣1 or anti-Ret antibodies. Immunoprecipitates were analyzed by immunoblotting with anti-GDNF, anti-Ret, and antiphosphotyrosine antibodies, respectively. The Western blots were shown in Figs. 8 -10, 10 and summarized in Table I. Combined mutants GFR␣1(N152A/N153A) and GFR␣1(S316A/N317A/ S318A) lost the ability to bind to GDNF and could not induce Ret phosphorylation but retained the ability to bind to Ret. Mutant GFR␣1(R259A) did not bind to GDNF or Ret and could not induce Ret phosphorylation. The mutant GFR␣1(Q247A/ D248A/S249A) retained the ability to bind to GDNF but had a significant reduction in ligand-stimulated Ret phosphorylation. The other GFR␣1 mutants (GFR␣1(R171A), GFR␣1(R197A), GFR␣1(K206A), GFR␣1(E270A), GFR␣1(D284A), GFR␣1-(L326A/K327A), GFR␣1(D334A), and GFR␣1(K339A)) retained binding capacity to GDNF and Ret and could induce normal levels of Ret phosphorylation as compared with wild type GFR␣1 (Figs. 8 -10). These results indicate that two clusters of residues, 152 NN 153 and 316 SNS 318 , in GFR␣1 are critical for GDNF binding and are indispensable for ligand-stimulated Ret phosphorylation. However, they are not involved in binding to Ret. The residue Arg 259 is essential in GDNF and Ret binding and Ret phosphorylation. Residues 247 QDS 249 appear not to be critical for binding to either GDNF or Ret but may influence the stability of the GDNF-GFR␣1-Ret complex and then influence ligand-stimulated Ret phosphorylation. The rest of the residues tested (Arg 171 , Arg 197 , Lys 206 , Glu 270 , Asp 284 , 326 LK 327 , Asp 334 , and Lys 339 ) appear not to be critical for GFR␣1 binding to GDNF and Ret.
The Effects of the Mutations on the Conformation and Stability of the Protein-In order to verify that the loss of activities of GFR␣1 mutants is due to changes in binding and not simply due to unfolding of the proteins, CD experiments and unfolding tests under different concentrations of GdmCl of these proteins were performed to assess the effects of the mutations on the structure and stability of the protein. The far-UV CD spectra of the four mutants (GFR␣1(N152A/N153A), GFR␣1(Q247A/ D248A/S249A), GFR␣1(R259A), and GFR␣1(S316A/N317A/ S318A)) are similar to that of wild type GFR␣1 (Fig. 11A), indicating that the secondary structures remain unchanged upon mutations. The CD-monitored unfolding curves of the four mutants also show no significant changes (Fig. 11B), indicating that the four mutations do not affect the unfolding properties of GFR␣1. DISCUSSION In this study, alanine substitution was applied to the trace residues of the GFR␣1 receptor with relatively high solvent accessibility to investigate their role in GDNF and Ret binding and subsequent receptor complex activation. The original evolutionary trace analysis is an all-or-none consensus sequencebased method that treats all columns with variable amino acid residues as non-conserved, regardless of the physicochemical similarity between them and thus may affect the sensitivity of this method (26). In this study, the improved ET method (29,30) was used to identify the functional epitopes in GFR␣ family members. Amino acid exchange matrices were used to better tolerate variations in each column, and each sequence was weighted according to its level of similarity with others to prevent the over-representation of similar sequences in the protein data base.
It has been reported that the central domain of GFR␣1 is a crucial determinant of ligand binding specificity and is critical for GDNF-induced neurotrophic function. The carboxyl-terminal segments adjacent to the central domain are necessary and have modulatory functions in ligand binding (25,36). However, which residues in central domain and carboxyl-terminal segments are responsible for its biological function were not known. Based on the evolutionary trace method, secondary structure, and relative solvent accessibility predictions of residues, 12 sites in rat GFR␣1 were selected for alanine mutagenesis. PC12 cells, which express low levels of endogenous Ret  (19,20), have been extensively used as a model system for peripheral neuronal differentiation (37). Each GFR␣1 mutant was co-transfected into PC12 cells with Ret cDNA to establish a series of GFR␣1-Ret PC12 cell lines.
GFR␣1 is an extracellular protein that is attached to the cell surface by GPI anchor (19,20). Co-expression of GFR␣1 mutants and Ret were verified by RT-PCR and Western blot (Figs. 5 and 6), and immunocytochemistry staining confirmed the right cellular distribution of all GFR␣1 mutants in these stable PC12 cells (Fig. 7). From differentiation and survival assays performed in these PC12 stable cell lines, we initially concluded that residues 152 NN 153 , 316 SNS 318 , and Arg 259 are critical for ligand binding and mediating the biological effects of GDNF on inducing neurite outgrowth and promoting survival (Figs. 2 and 3). Residues 247 QDS 249 , when mutated to alanine, had reduced differentiation and survival response of PC12 cells to GDNF, as compared with wild type GFR␣1. This finding indicates that these three amino acids are not critical for ligand binding but may have modulatory function and influence the stability of the multisubunit receptor complexes (Figs. 2 and 3). Later, the immunoprecipitation and Western blot results further confirmed that mutants GFR␣1(N152A/N153A) and GFR␣1(S316A/N317A/S318A) lost the capability to bind to GDNF but could still bind to Ret (Figs. 8 and 9). This finding suggests that there are at least two distinct binding sites in GFR␣1 for GDNF and Ret binding, respectively. We speculate that GFR␣1 might pre-associate weakly with Ret in the absence of ligand, and GDNF stimulation enhances the GDNF-GFR␣1-Ret complex formation. In our study, the GFR␣1(N152A/N153A) mutation retained Ret binding capacity, which was consistent with the conclusion of a previous study (25) that the first 161 residues from the amino terminus of GFR␣1 might be deleted without affecting the ability of GFR␣1 to interact with Ret (25). These findings suggest that GFR␣1 does have an interface directly interacting with Ret.
Arg 259 in GFR␣1 is crucial not only for GFR␣1 and Ret binding but also for activation of the receptor complex. This residue is a globally invariant residue that is completely conserved in the entire GFR␣ family, suggesting a similar role for this site in all GFR␣ members. That a single mutation R259A in GFR␣1 lost its ability to bind both GDNF and Ret suggests that a point mutation in one subunit of heterotrimeric complex could lead to simultaneous loss of binding with the other two subunits. A comparable example for this situation is Lis1. A point mutation of H149R or S169P in Lis1 led to loss of binding to either PAFAH1B2 or PAFAH1B3 subunit and impairment in the formation of platelet-activating factor acetylhydrolase (PAFAH1B) (38). The heterotrimer, PAFAH1B complex, is required in the process of neuronal migration during brain development. The above two point mutations in human Lis1 result in the lissencephaly phenotype (38,39). These phenomena imply that Arg 259 in GFR␣1, which is crucial for both GFR␣1 and Ret binding, may also cause clinical symptoms upon mutation in vivo.
The far-UV CD spectra and unfolding experiments show that the secondary structures and protein stabilities of the GFR␣1 mutants are not significantly changed compared with wild type GFR␣1. So it was verified that loss of activities of GFR␣1 mutants are due to changes in binding and not simply due to unfolding of the proteins.
We demonstrated that combined residues 152 NN 153 and 316 SNS 318 of GFR␣1 are crucial for GDNF binding and mediating neurotrophic function of GDNF. Later individual alanine mutation of each of these residues was transfected into PC12 cells, respectively. Mutants N152A, N153A, S316A, and S318A had a major effect on GDNF-induced neurite outgrowth, indicating that they make important contributions to GDNF binding. However, N317A did not alter the response of GDNF in these PC12 cells, suggesting that it is not critical for GDNF-GFR␣1 complex formation (data not shown).
In this study, we have identified a set of key amino acids in the GFR␣1 receptor to be critical for functional receptor complex formation and eliciting downstream signal transduction. These residues may contribute synergistically to GDNF binding and form a functional epitope in GFR␣1. Other sites that may also take part in the formation of the functional epitope of this molecule are still under investigation. Several GFR␣1 mutants deficient in GDNF binding are still able to bind with Ret. This finding confirms that there are two distinct structural determinants in GFR␣1 for GDNF binding and association with Ret, respectively. But a single point mutation in GFR␣1 simultaneously lost its ability to bind with GDNF, and Ret implies that the two binding sites may have some overlap. FIG. 11. Circular dichroic spectra and denaturation curves of wild type GFR␣1 and its mutants. A, the far-UV CD spectra of all mutants were similar to that of the wild type GFR␣1, indicating that the secondary structures remained unchanged by mutations. B, GdmClinduced unfolding experiment was followed by monitoring CD ellipticity change at 220 nm. The unfolding curves of all mutants also showed no significant changes compared with wild type GFR␣1, indicating that the protein stability of GFR␣1 mutants and wild type GFR␣1 was almost same.