The Journal of Neuroscience, March 1993, 13(3): 941-951

HSV-1 Vector-Mediated Gene Transfer of the Human Nerve Receptor ~75 hNGFRDefines High-Affinity NGF Binding

David S. Battleman,’ Alfred I. Geller,* and Moses V. Chaol ‘Department of Cell Biology and Anatomy, Cornell University Medical College, New York, New York 10021 and *Division of Endocrinology, Children’s Hospital, Boston, Massachusetts 02115

A series of recombinant herpes simplex virus (HSV-1) vec- signaling mechanism for these factors has been defined with the tors have been constructed that encode either the full-length discovery that the p140proto1rkreceptor (trk) in- cDNA of the human p75 NGF receptor (~75~~9 or truncated teracts specifically with NGF, resulting in increased tyrosine forms of the receptor. Infection of cultured fibroblast cells phosphorylation of cellular substrates in PC1 2 cells (Kaplan et with viral stocks results in abundant expression of all three al., 1991 b; Klein et al., 1991a). Other trk family members, in- cDNAs, as detected by affinity cross-linking, immunoblot cluding trkB and trkC, have since been found to interact with analysis, and equilibrium binding. Furthermore, viral infec- multiple neurotrophic ligands (Cordon-Card0 et al., 199 1; Klein tion of primary neuronal cultures gives easily detectable p75 et al., 199 la,b; Lambelle et al., 199 1; Soppet et al., 199 1; Squint0 expression by immunofluorescence and affinity cross-link- et al., 1991). ing. When p75 was introduced by viral infection into fibro- An important question concerning the action of neurotro- blast cells expressing the frkproto-oncogene, a new binding phins is how the functional receptor complex is encoded and site was created, consistent with high-affinity NGF binding. specified. NGF and the other family members are This site is not created by the coexpression of truncated known to interact with a 75,000 Da transmembrane receptor, forms of p75 that lack either the extracellular binding ~75~~~~ (Johnson et al., 1986; Radeke et al., 1987; Emfors et domain or the cytoplasmic domain of the receptor, sug- al., 1990; Rodriquez-Tebar et al., 1990; Squint0 et al., 1991). gesting that both of these regions of the receptor are re- NGF binds to this receptor with low affinity (Kd = 1O-9 M), but quired for the formation of the high-affinity NGF binding site. this interaction does not initiate any measurable signal trans- Hence, these HSV-1 vectors give rise to appropriate NGF duction in fibroblast cells expressing ~75~~~~ (Hempstead et al., receptor binding after viral infection. The application of these 1989). Only in NGF-responsive cells are high-affinity (Z& = HSV-1 constructs to primary neuronal culture and in vivo lo-” M) and low-affinity binding sites displayed (Sutter et al., models of ~75~~~~ function is discussed. 1979; Schechter and Bothwell, 198 1; Green and Greene, 1986). [Key words: NGF receptor, herpes simplex virus, primary High-affinity binding is required for functional responses to NGF culture, trk oncogene, neurotrophin, receptor binding] both in culture and in vivo and can be reconstituted in cultured cells that express p140prototrk(Hempstead et al., 1991; Klein et NGF results in cell survival and mainte- al., 1991a). Whether this binding requires the concomitant ex- nance of discrete neuronal populations, including sensory neu- pression of p7YGFR has been the subject of debate (Bothwell, rons of neural crest origin and sympathetic (Thoenen 199 1; Chao, 1992). and Barde, 1980; Levi-Montalcini, 1987). NGF also regulates Both receptors are known to be colocalized in sensory, sym- neural populations in the brain, most notably cholinergic fibers pathetic, and central cholinergic neurons during NGF-respon- of the (Gnahn et al., 1983; Hefti, 1986; Mobley sive periods (Buck et al., 1987; Hohaman et al., 1992; Schech- et al., 1986). The identification of BDNF, terson and Bothwell, 1992), and monoclonal antibodies against NT-3, NT-4, and NT-5 has uncovered a family of related poly- ~75 have been shown to display inhibitory effects upon peptide differentiation factors with similar biological effects regeneration (Chandler et al., 1984) and c-fos induction in PC1 2 (Leibrock et al., 1989; Emfors et al., 1990; Hohn et al., 1990; cells (Milbrandt, 1986). Previous equilibrium binding studies Maisonpierre et al., 1990; Rosenthal et al., 1990; Berkemeier have indicated that ~75~~~~ participates in high-affinity NGF et al., 1991; Hallbook et al., 1991). Furthermore, a primary binding(Green and Greene, 1986; Hempstead et al., 1989, 199 1; Received Apr. 24, 1992; revised July 27, 1992; accepted Sept. 2, 1992. Pleasure et al., 1990; Matsushima and Bogenmann, 199 1) and We are grateful to Drs. Howard Federoff and’Roy Silverstein for critical reading that ~14.0~~~~~‘~~binds to NGF with low affinity (Kaplan et al., of the manuscript. We thank Dr. Barbara Hempstead and Dr. Margaret Berg for 199 1a). Other experiments, however, have reached the opposite many helpful discussions and technical assistance. Also, we thank both Dr. Nila conclusion, suggesting that the unique expression of p140pro101rk Patil and Margret Einarson for use of the NGF receptor mutant constructs and Dr. Luis Parada for the trk-3T3 cells. This work was supported by grants from defines a high-affinity receptor for NGF (Klein et al., 1991a). the National Institutes of Health, the American Cancer Society, the Zenith Award The involvement of ~75~~~~ in NGF signal transduction was from the Alzheimer’s Association, and the Dorothy Rodbell Cohen Foundation recently discounted with results demonstrating that p 140prolo1rk to M.V.C. A.I.G. was supported by the American Health Assistance Foundation, the American Parkinson’s Disease Association and Alkermes Inc. A.I.G. thanks and the other tyrosine kinase family members could Oliver Ashe for technical assistance. function independently in fibroblast cells without the coexpres- Correspondence should be addressed to Dr. Moses Chao, The Department of Cell Biology and Anatomy, Cornell University Medical College, 1300 York Av- sion of ~75~~~~ (Cordon-Card0 et al., 199 1; Glass et al., 199 1). enue, New York, NY 1002 1. Moreover, use of anti-p75NCFR antibodies and site-directed mu- Copyright 0 1993 Society for Neuroscience 0270-6474/93/130941-l 1$05.00/O tants of NGF has implied that the biological effects of NGF 942 Battleman et al. + p7PNGFR Defines High-Affinity NGF Binding

may be mediated through the trk receptor alone, without binding virus subsequently harvested, passaged, and titered (8 x lo5 infectious to p7YGFR (Weskamp and Reichardt, 199 1; IbZnez et al., 1992). particles/ml) as described previously (Miller and Hyman, 1978). Northern blot analysis. Cells were grown to 70% confluence on 100 To address the functional role of the ~75~~~~ in mediating mm tissue culture dishes, infected with 0.4 ml of viral stock at a mul- the differentiative effects of NGF action, we have sought to tinlicitv of infection (MOB of 0.25. and incubated for 6-24 hr at 37°C. express_ _~75 cDNAs in defective herpes simplex virus (HSV-1) and total mRNA was prenared usinn the RNASol B method (Chom: vectors (Geller and Breakefield, 1988; Geller et al., 199 1). Here czynski and Nicoletta,-1987). For Northern blot analysis, mRNA (20 &sample) was separated by electrophoresis in 1% agarose gels con- we report the construction and expression of three novel HSV- 1 taining 2.2 M formaldehyde, transferred to nitrocellulose filters, and vectors carrying a full-length human ~75 receptor cDNA as well baked in vacua at 80°C for 4 hr (Sambrook et al., 1989). The presence as two mutant receptor clones, each lacking an essential domain of mRNA species was confirmed by ethidium bromide visualization. of the receptor. Our results confirm that the coexpression of Filters were then prehybridized in 50% formamide, 5 x SSC, and 5 x both p75NoFR and ~140~~~‘~~~~ is required for the functional re- Denhardt’s solution with 500 &ml of denatured salmon sperm DNA at 42°C for at least 4 hr. Hybridization with radiolabeled probe was constitution of a high-affinity binding site for NGF. The ex- carried out at 42°C in 50% formamide, 5 x SSC, and 1 x Denhardt’s pression of trk with either the ligand binding domain or the solution with 100 pg of denatured salmon sperm DNA. Randomly cytoplasmic domain of p75, however, was not sufficient to me- primed restriction fragments were generated to specific activity of 5 x diate this high-affinity interaction. Moreover, expression studies lo8 cpm/pg. Filters were washed to 0.4 x SSC at 68°C and exposed to x-ray film at -70°C with an intensifying screen. in primary culture indicate that the HSV-1 vectors employed Immunoblotting. For immunoblot analysis, infected cell cultures were in this study are capable of faithfully delivering high levels of prepared as above. Infected fibroblasts were washed in phosphate-buf- receptors into neurons, and therefore these HSV- 1 vectors will fered saline (PBS; pH 6.4) and rapidly solubilized in 1 ml of solubili- be useful in addressing questions about the physiological roles zation buffer [lo mM Tris-HCl (pH 7.4) 1% SDS, and 0.1 mM each of phenylmethylsulfonyl fluoride (PMSF)/aprotinin/leupeptin]. Cell lys- plays in nerve cells. that ~75~~~~ ates were incubated on ice for 15 min and then pelleted at top speed in a microfuge at 4°C for 10 min. Supematants were assayed for concentrations with a protein assay from Bio-Rad (Richmond, CA). Materials and Methods (200 pg) were immunoprecipitated with anti-p75 antibody, electrophoresed by SDS-PAGE (8%), and electrophoretically transferred Materials. Restriction endonucleases were purchased from New England to nitrocellulose. Filters were then incubated with anti-p75 antisera Biolabs (Beverly, MA), Mouse NGF (p-dimer) was obtained from Bio- followed by lZ51-labeled protein A from Amersham (Arlington Heights, products for Science (Indianapolis, IN). Anti-2.5s NGF antiserum, a IL). Autoradiography was carried out at -70°C with an intensifying rabbit anti-mouse polyclonal antibody, was purchased from Collabo- screen. rative Research (Bedford, MA), and anti-p75 antibody, a polyclonal Affinity cross-linking and immunoprecipitation. 3T3-trk cells infected antibody that was generated against the cytoplasmic domain of the low- at an MO1 of 0.25, and uninfected control cells were harvested for affinity NGF receptor, was generously supplied by Dr. S. J. Decker affinity cross-linking. NGF receptors were labeled by affinity cross-link- (Parke-Davis, Ann Arbor, MI). Antibody ME20.4 against the human ing methods previously described (Green and Greene, 1986; Hempstead NGF receptor (hNGFR; Ross et al., 1984) was harvested from cultured et al., 1989). For EDAC cross-linking to ‘Y-NGF, cells were washed hybridoma cell line supematants (from American Type Culture Collec- in PBS (pH 6.5) and resuspended to a final concentration of 2 x lo6 tion), and rhodamine-conjugated goat anti-mouse (Fab’), antibody was cells/ml in PBS containing 1 mg/ml each of glucose and bovine serum purchased from Jackson Labs (Bar Harbor, ME). Ethyl-3-(dimethylam- albumin. lZ51-NGF was then added to each cell suspension at a final inopropyl) carbodiimide (EDAC) was purchased from Pierce Chemical concentration of 25 r&ml, in the presence or absence of 5 wt~ unlabeled Co. (Rockford, IL). lZ51-NGF was prepared by the lactoperoxidase meth- NGF for 2 hr on a rotator at 4°C. EDAC was next added to a final con- od as described previously (Green and Greene, 1986). centration of 4 mM at room temperature for 30 min. Alter cross-linking, CelI culture. PC 12 cells were maintained in Dulbecco’s modified Ea- cells were centrifuged, washed in PBS-lysine (50 mM), andsolubilized gle’s medium (DMEM) with 10% fetal bovine serum and 5% horse in lysis buffer (10 mM Tris, pH 7.6, 66 mM EDTA, 1% NonidetP40, serum (GIBCO, GrandIsland, NY). NIH-3T3 cells, which overexpress 0.4% deoxycholate, and 0.1 mM each of PMSF/aprotinin/leupeptin) at ~140~~~‘~~~~ (3T3-trk cells) (Kaplan et al.. 1991a). were maintained in 4°C for 20 min. Lysates were immunoprecipitated with anti-2.5s NGF DMEM with 10% fetal bovineserum and 0.4 mg&tl of G4 18 (GIBCO). antisera at 4°C for 2 hr and washed extensively in RIPA buffer (10 mM Construction of vectors and packaging of HSV-I virus particles. ~75 Tris-HCl, pH 8.0, 150 mM NaCl, 10 mM KCl, 1% deoxycholate, 1% human NGF receptors pSL and pxl were constructed as described pre- Nonidet P-40, 0.1% SDS, and 1 mM EDTA). The samples were then viously (Hempstead et al., 1990). For mutant p5A, a 1.5 kilobase (kb) resuspended in sample buffer, boiled for 5 min, and subjected to elec- hNGFR cDNA was digested with both StuZ and Sau3A and religated. trophoresis on a 7.5Oh SDS-polyacrylamide gel. The dried gels were The resulting 1 .O kb fragment lacking nucleotides 2 15-682 was recov- exposed to autoradiographic film at - 70°C with an intensifying screen. ered. The pSL, pxl, and p5A cDNA inserts were then isolated from Equilibrium binding analysis. Infections were carried out as described carrier sequence by EcoRI and subcloned into the recipient above. Equilibrium binding of 12’I-NGF to crude cellular membrane HSV-1 vector pHSVPrpUC (Geller and Breakefield, 1988). pHSV- preparations was performed as described previously (Hempstead et al., hNGFR. DHSV-xl, and DHSV-5A contain the DSL insert (2.3 kb), the 1989). All measurements were verified in trinlicate._ I and the data were px 1 insert (2.3 kb),’ and the p5A insert (1 .O kb), respectively. The ori- assessed by Scatchard analysis (Scatchard, 1949). Only binding values entation of each insert was verified by restriction enzyme analysis. above 50% specific binding were used in the final analysis. The LIGAND pHSVlac contains the 1acZ gene and was constructed as previously program was used to analyze the data (Munson and Rodbard, 1980). described (Geller and Breakefield, 1988). Primary neuronal cultures and immunohistochemistry. Timed-prea- The packaging system for HSV- 1 viral particles has been previously nant Sprague-Dawley rats [embryonic day 16 (E 16)] were purchased reported (Spaete and Frenkel, 1982; Geller, 1988). In brief, 1.5 x 10s from Charles River Labs (Wilmington, MA) and embryos removed CVl cells were plated on 60 mm dishes and transfected with 0.5 ml of under sterile conditions. Brain tissue was isolated and the cerebral cortex a calcium phosphate precipitate containing 1 pg of pHSV- 1 vector DNA was carefully dissected away from hippocampus, basal forebrain, and and 9 pg of salmon sperm DNA. Four hours later the cells were treated olfactory bulbs. Meninges were also removed. Cortical cells were then with 15% glycerol (Parker and Stark, 1987) and incubated at 37°C for manually dissociated and plated on poly-D-lysine-coated dishes. Cul- 24 hr. Cells were then infected with 1.5 x 1O6 plaque-forming units of tures were enriched for nerve cells by maintaining cultures in serum- HSV-1 strain 17 ts K (Davison et al., 1984) in 100 ~1 of medium. HSV- 1 free conditions for at least 5 d as described previously (di Porzio et al., ts K is a mutant HSV-1 strain that contains a temperature-sensitive 1980). mutation in the IE3 gene, resulting in an immediate-early phenotype For immunohistochemistry, primary cortical cells were infected at an and no viral replication at the restrictive temperature of 37°C. After 1 MO1 of 0.25 for 8 hr at 37°C. Fresh media was then added and the cells hr at room temperature, an additional 5 ml of medium were added to were placed at 37°C for an additional 36 hr mior to fixation. Cells were each infected plate. Cultures were then incubated for 72 hr at 3 1°C and fixed -in buffered 4% paraformaldehyde for 1 hr on ice, washed with The Journal of Neuroscience, March 1993, U(3) 943

Figure I. Schematic representation of p75hNGFR and receptor mutants. Con- structs were generated as outlined in the Materials and Methods. The hNGFR Construct Cyeteine Repeat Domains serine - Threonine TM Cytoplasmic Domain construct represents the full-length R&Xl cDNA of the human low-affinity NGF bNGFR receptor, ~75. Mutant-xl encodes for an intracellular deletion receptor that is truncated 4 amino acids after the trans- membrane domain at amino acid DO- Mutant-xl sition 248. Mutant-5A encodes for an extracellular deletion mutation that lacks the NGF binding domain due to a large deletion (Welcher et al., 199 1; Mutant-SA Yan and Chao, 199 1) spanning amino NH2b ‘OOH acids 34-191.

PBS, and preincubated in PB buffer (0.1 M PBS, pH 7.4, with 0.1% defective HSV-1 vector (Geller and Breakefield, 1988), which Triton X-100 and 1 mg/ml BSA) for 1 hr at room temperature. ME20.4 is shown in Figure 2, and then packaged into HSV-1 particles hybridoma supematants were diluted 1:5 in PB buffer and incubated as outlined in the Materials and Methods. with cells overnight at 4°C. Cells were thoroughly washed and then incubated in a 1:250 dilution of a rhodamine-conjugated goat anti- Expression studies in cell culture mouse secondary antibody (Jackson Labs). Cells were washed exten- To determine if the HSV- 1 vectors appropriately express human sively and mounted. Control cultures were assayed in the absence of primary antibody. P75 NGFR (P75hNGFR ) and mutant receptor constructs, 3T3-t&cells were infected with 8 x lo5 infectious particles/ml of viral stock and analyzed for both mRNA and cell surface protein. These cells were chosen as the recipient cell line because they lack Results endogenous p7YGFR (Kaplan et al., 199 1a; Klein et al., 199 la). hNGFR constructs and HS V- 1 vectors Expression of all three receptor constructs in fibroblasts is readi- The NGF receptor constructs used in this study are shown sche- ly detectable by Northern blot analysis (Fig. 3A). In Figure 3A, matically in Figure 1. The hNGFR construct is a 2.3 kb full- a prominent 2.3 kb message was detected in pHSV-hNGFR and length cDNA of the human NGF receptor that contains the pHSV-xl lanes (lanes 2 and 3) while in mock-infected 3T3-trk entire coding region of the receptor but lacks most of the 3’- cells, no signal was detected (lane 1). untranslated sequences. The cytoplasmic deletion mutant Cells infected with pHSV-5A also expressed a message of hNGFR-xl was generated by the site-directed insertion of a about 2.3 kb in size (Fig. 3A, lane 4). Though the cDNA insert universal Xba-termination linker 4 amino acids beyond the of this construct was approximately 1.0 kb, the increased size transmembrane domain at amino acid position 248 (Hempstead of this primary transcript was likely due to an alternative po- et al., 1990). The extracellular deletion mutant hNGFR-SA con- lyadenylation event. Unlike the pHSV-hNGFR and pHSV-xl tains a large deletion in the ligand binding domain (Welcher et constructs, which contain an endogenous polyadenylation signal al., 199 1; Yan and Chao, 199 I), spanning amino acids 34-l 9 1. in their cDNA insert, the pHSV-5A construct does not. There- Each of these receptor constructs was subcloned into the 4.8 kb fore, pHSV-5A must rely on the SV40 polyadenylation signal

HSV-1 IE 415 Figure 2. Structure of pHSV-hNGFR. This HSV-1 expression vector is a 4.8 kb plasmid that contains the 2.3 kb p75NGFR hNGFR insert and the following set of gene genetic elements: (1) sequences from pBR322 that allow for bacterial repli- cation and selection of the plasmid, in- cluding the ColEl origin of DNA rep- lication and the ampicillin-resistance gene; (2) sequences from the HSV- 1 ge- nome that allow for packaging of the plasmid into viral particles, including the HSV-1 packaging site and herpes simplex virus DNA replication ori, (Davison and Wilkie, 198 1; McGeoch SW0 et al., 1986); and (3) sequences from ~01~ A site both HSV-1 and SV40 to compose el- ements of the transcription unit, in- cluding the HSV-1 IE 4/5 promoter (McGeoch et al., 1986) and the SV40 early region polyA site (Hall et al., 1983). 944 Battleman et al. l ~75”~~~~ Defines High-Affinity NGF Binding

A. Northern Blot B. lmmunoblot

+ IgG 28s

18S 28

C. Affinity Crosslinking PC12 3T3- trk 3T3- trk 3T3- trk 3T3- trk uninfected Mock HSV-5A HSV-xl HSV-hNGFR -- 1--

I + - +-+-+-+

228

Figure 3. Expression of p7YNGFR and receptor mutants in 3T3-trk cells using recombinant HSV-1 vectors. A, Northern blot analysis of RNA isolated from 3T3-trk cells infected with HSV- 1 viral stocks. Cells were infected for 6-24 hr at 37”C, and RNA was isolated 36 hr later as described in Materials and Methods. Northern blots were probed with radiolabeled B9 fragment, a probe specific to ~75”~~~ cDNA. Each lane contained 20 pg of total RNA. Lane I is RNA harvested from fibroblasts following mock infection. Lanes 2-4 contain RNA from fibroblasts infected with 0.4 ml ofHSV-hNGFR, HSV-x 1, and HSV-SA, respectively. pHSVlac virus, a vector that expresses the E. coli @galactosidase (Geller and Breakefield, 1988) was used for control infection. Positions of 28s and 18s ribosomal markers are indicated. B, Immunoblot analysis with anti-p75 antibody of HSV-l-infected 3T3-trk cells. Proteins (200 fig) from infected cells were immunoprecipitated and blotted with anti-p75 antibody as described in Materials and Methods. Anti-p75 antibody recognizes the cytoplasmic domain of human ~75. Lane I represents mock-infected 3T3-trk cells. Lanes 2-4 are 3T3-trk cells infected with HSV-hNGFR, HSV-x 1, and HSV-SA, respectively. Molecular mass markers (in kilodaltons) are indicated. An unlabeled arrow indicates the position of the HSV-5A protein at about 35 kDa. The broad IgG band (arrow) runs at approximately 50 kDa. pHSVlac virus was used for control infection (Geller and Breakefield, 1988). C, Affinity cross-linking of NGF receptors in cultured cell lines. NGF receptors were cross-linked to 1251-NGF using EDAC in the presence (+) or absence (-) of excess unlabeled NGF as outlined in Materials and Methods. Samples were loaded as indicated above the lanes. pHSVlac virus was used for control infection (Geller and Breakefield, 1988). Molecular mass markers (in kilodaltons) are indicated on the left. Arrows indicate the position of cross-linked ~75 at 100 kDa and the cross-linked HSV-x 1 protein at approximately 68 kDa. that is present in the HSV- 1 vector about 1.15 kb downstream with either the full-length receptor HSV-hNGFR or the intra- of the cloning site (Hall et al., 1983). cellular deletion mutant HSV-xl (Fig. 3B, lanes 2,3). The broad To verify further that the HSV- l-infected cells appropriately IgG band is observed across the top of each lane at approxi- expressed these cDNAs, we assayed both for cell surface protein mately 50 kDa. expression and for ligand binding. Figure 3B shows an immu- To detect the hNGFR and mutant-xl proteins, an affinity noblot of proteins immunoprecipitated with anti-p75 antibody. cross-linking assay was used (Fig. 3C). This cross-linking re- The polyclonal anti-p75 antibody was raised against the cyto- action results in a 90-100 kDa labeled species in PC12 cells, plasmic domain of ~75~~~~ and was used to confirm the ex- human melanoma cells, and rat brain, and has been shown to pression ofthe pHSV-5A construct. Cells infected with the HSV- represent a complex between labeled NGF and the ~75 receptor 5A virus demonstrated a major 35 kDa protein of predicted (Grob et al., 1983; Green and Greene, 1986). Infected 3T3-trk size (Fig. 3B, lane 4). No signal was detected in this range in fibroblast and PC1 2 cells were incubated with lZ51-NGF in sus- mock-infected fibroblasts (Fig. 3B, lane 1) or in cells infected pension for 2 hr, cross-linked with EDAC, and then immuno- The Journal of Neuroscience, March 1993, U(3) 945

500 loo0 Iyw) m 2m 3mm : m Bound (fmoWmg) Bound (fmoles/mg)

5.3lStrk I HSV-xl I

I ,‘.‘.,‘~‘.l....l.‘..l...~ sw Bouri?(fmole%ng) lmo ’ Figure 4. Equilibrium binding and Scatchard plot analysis of NGF receptors in PC 12 cells and HSV- l-infected 3T3-trk cells. 3T3-trk cells were infected with viral stocks for 6-24 hr and crude membrane preparations were prepared 36 hr later as described previously (Hempstead et al., 1989). Binding assayswere performed as presented in the Materials and Methods. Binding data were graphed according to Scatchard (1949). A is the Scatchard plot of uninfected PC 12 cells. B is the Scatchard plot of uninfected 3T3-trk cells. C is the Scatchard plot of mock-infected 3T3-trk cells. D-F show the binding curves for 3T3-trk cells infected with the HSV-hNGFB, HSV-xl, and HSV-SA viruses, respectively. Control infections were performed using pHSVlac (Geller and Breakefield, 1988). precipitated with anti-NGF antibody (see Materials and Meth- been introduced has also been observed (B. L. Hempstead, un- ods). 3T3-trk cells infected with HSV-hNGFR viral stock published observations). The significance of this event is pres- demonstrated the same cross-linked complex as PC1 2 cells (100 ently

Table 1. Summary of Scatchard plot analysis

Affinity site 1 Affinity site 2 Site 1: Cell line Infection Temperature Kd (M) Sites/mg of protein Kd (M) Sites/mg of protein Site 2 PC12 Uninfected 30°C 4.15 x 10-P 8.54 x 10”’ 6.62 x lo-” 4.93 x 10”” 17:l 3T3-trk Uninfected 30°C 4.23 x 1O-9 4.53 x lo”2 N.A. 3T3-trk pHSVIac 30°C 5.61 x lo+ 6.76 x lo+‘* N.A. 3T3-trk HSV-hNGFR 30°C 3.99 x 10-g 4.69 x lo+‘* 5.11 x 10-I’ 1.44 x 10”’ 32:l 3T3-trk HSV-x 1 30°C 4.81 x 10e9 3.12 x lo+‘* N.A. 3T3-trk HSV-5A 30°C 10.46 x lo+’ 5.53 x lo+‘2 N.A.

All binding data presented above were analyzed using the WGAND program (Munson and Rodbard, 1980). N.A. indicates that only one kinetic site was identified. Temperatures refer to the equilibrium binding assay conditions.

to determine if we could reconstitute high-affinity NGF binding curve, with a Kd of 5.67 x 1O-9 M (Fig. 4C). In contrast, when in 3T3-trk cells, a recipient fibroblast cell line that overexpresses membranes were assayed from 3T3-trk cells infected with HSV- the trk proto-oncogene at high levels (> 200,000 sites per cell). hNGFR, a biphasic Scatchard plot was observed, consisting of Since the high-affinity NGF binding complex can be distin- two discrete binding sites (Fig. 40). Similar to PC12 cells, the guished from low-affinity complex by equilibrium binding, Scat- reconstituted high-affinity site represents a small percentage of chard analysis was conducted. A membrane binding assay was the total number of NGF binding sites and has a Kd of 5.11 x employed in order to minimize the effects of internalization of 10-l’ M. The low-affinity site in the HSV-hNGFR-infected cells NGF. Crude cell membranes were prepared from PC1 2 cells has a Kd of 3.99 x 1O-9 M. A summary of these results is shown and HSV- l-infected 3T3-trk cells. PC 12 cell membranes (Fig. in Table 1. From these data, we conclude that infection of 3T3- 4A) displayed two binding sites for NGF as previously reported trk cells with HSV-hNGFR is capable of creating a new binding (Landreth and Shooter, 1980; Block and Bothwell, 1983). The site for NGF, and therefore p7SNGFRappears to be required in high-affinity binding site represents only about 5% of the total the high-affinity NGF binding complex. number of NGF binding sites and demonstrates a Kd of 6.62 x To investigate the nature of the high-affinity binding site more lO-L* M. The low-affinity site has a Kd of 4.15 x 1O-9 M. completely, we sought to express and characterize the binding Multiple binding studies of the 3T3-trk recipient cell line properties of our mutant constructs of p75 in the presence of demonstrate that this cell line contains predominately low-af- pl 40pr0M’rk.In these experiments, 3T3-trkcells were infected with finity receptors for NGF. A representative Scatchard plot is either the HSV-xl or the HSV-SA virus and assayed by Scat- shown in Figure 4B. A single binding site with a Kd of 4.23 x chard analysis as above. Representative binding curves are shown 10-P M is observed, consistent with a low-affinity interaction. in Figure 4, E and F. In both cases, linear binding curves were pHSVlac infections of 3T3-trk cells generated a similar binding generated with a Kd of 4.81 x 1O-9 M for HSV-xl and a Kd of

PC12 3T3-frk Cortex Cortex uninfected Mock Mock HSV-hNGFR I-I- + - + - + - +

Figure 5. EDAC affinity cross-linking of NGF receptors in primary cortical neurons infected with HSV-hNGFR. 200 Primary cortical cultures were dissected from El6 rat brain and cultured for 5 d in serum-free media as indicated in Materials and Methods. Neuronal cul- tures were then infected with either pHSVlac virus or HSV-hNGFR virus for 6 hr. Following this incubation pe- riod fresh medium was added, and cells 97 were harvested 36 hr later. Uninfected PC1 2 cells and mock-infected 3T3-trk cells were used as controls. NGF recep- tors were labeled by the cross-linking 68 to Y-NGF as described in Figure 3C. Lanes are loaded as indicated at the top, and all samples contain 2 x lo6 cells. Molecular weight standards (in kilo- 43 daltons) are shown on the left. The Journal of Neuroscience, March 1993, 13(3) 947

Fzgure 6. Immunofluorescent localization of hNGFR in primary cortical neurons infected with HSV-hNGFR. Primary cultures of cortical neurons were prepared as described in Figure 5. Following infection, cells were fixed in 4% paraformaldehyde for 1 hr at room temperature, washed extensively with PBS, and incubated overnight at 4°C with ME20.4 antibody. Cells were again thoroughly washed and then stained with a 1:250 dilution of a rhodamine-conjugated goat anti-mouse antibody (Jackson Labs). D demonstrates the imrnunoreactivity of cortical cultures infected with HSV-hNGFR. C is the corresponding bright field. B demonstrates the immunoreactivity of cultures infected with pHSVlac. A is the corre- sponding bright field. Note that the ~75”~~~~ is highly expressed both in the nerve cell body as well as in the distal processes.

10.46 x 10 9 M for HSV-SA. This binding profile is quantita- incubated with rZSI-NGF and chemically cross-linked with EDAC tively similar to that of membranes from control cells (Fig. as before. In comparison to endogenous levels of p75 in PC 12 4&C). Thus, overexpression ofeither the ligand binding domain cells, the cells infected with HSV-hNGFR were readily detected or the cytoplasmic domain of ~75 in the presence of high levels at high levels (Fig. 5). The addition of excess cold NGF in of p 140~~~‘~~~~gives rise to only low-affinity binding of rZ51-NGF, alternate lanes confirmed the specificity ofthe cross-linked prod- suggesting that neither of these two domains of ~75, when in- uct. Furthermore, in uninfected cultures, cross-linking was not dependently expressed, is sufficient for the functional reconsti- detected, suggesting that NGF receptors were either absent or tution of the high-affinity NGF binding site. at very low levels in E 16 cortical neurons. For immunohistochemistry, cultures were fixed in parafor- Expression of hNGFR in primary neuronal cultures maldehyde, infected with either pHSVlac or HSV-hNGFR vi- To demonstrate that HSV- 1 vectors could be used efficiently to rus, and then stained with ME20.4 antibody, a monoclonal introduce ~75 receptors into neurons, primary cortical cultures antibody specific for the human p75 receptor (Ross et al., 1984). were employed. Cultures were maintained in serum-free con- Representative immunofluorescent photomicrographs of these ditions for at least 5 d. These culture conditions have been cortical nerve cells are shown in Figure 6. In Figure 60, a brightly shown to enrich for nerve cells in the absence of mitotic inhib- stained nerve cell body with its less intcnscly stained processes itors (di Porzio et al., 1980). Neuronal cultures were infected centers the field. The corresponding phase-contrast photomi- with HSV-hNGFR virus or pHSVlac virus at 37°C. Following crograph reveals a field of about 20 cells and is shown in Figure infection, cells were either harvested for affinity cross-linking 6C. About 20-25% of these surrounding cells are expressing experiments or fixed in 4% paraformaldehyde for immunohis- lower levels of human receptor that could be detected above tochemistry. background. Figure 6A is the phase-contrast field ofcortical cells An equal number of PC12 cells and cortical neurons were infected with the pHSVlac virus. The corresponding immuno- 848 Battleman et al. * p7PNGFR Defines High-Affinity NGF Binding fluorescent field is shown in Figure 6B. These results indicate Therefore, the percentage of reconstituted high-affinity binding that HSV-mediated expression of ~75~~~~ can be readily ex- sites per cell may be significantly underestimated. For this rea- tended to primary nerve cell cultures. son, the relative number of high- and low-affinity sites for each cell line was reported as sites per milligram of protein (Table Discussion 1). A central theme that has emerged from the study of NGF is Scatchard analysis of 3T3-trk cells infected with either HSV- that a limited supply of neurotrophic factor is available for xl or HSV-5A has served to characterize further the nature of developing neurons during target innervation (Levi-Montalcini, the high-affinity NGF binding site. When either of these two 1987; Barde, 1989; Oppenheim, 1989; Davies, 1991). Com- constructs were expressed to high levels in the presence of petition for target-derived factors therefore depends upon the p140proto’rk, only low-affinity NGF binding was observed, sug- precise interaction of for specific receptors. For gesting that an intact ~75 receptor is required to form the high- NGF, two receptors have been identified, p7SNGFRand the prod- affinity complex. We can also conclude from these experiments uct of the trk proto-oncogene, p 140 protofrk.The colocalization of that ~75~~~~ is not merely functioning as a binding protein for both of these receptors in the majority of NGF-responsive cells NGF. If p75 served to present NGF to the high-affinity receptor, in viva suggests that the specificity of NGF action must be de- then the HSV-x 1 construct, which binds NGF, would have been rived from an interaction of NGF with both receptors. Fur- predicted to reconstitute a high-affinity binding site. The extra- thermore, the equilibrium binding constants of NGF binding cellular domain of ~75 alone is not sufficient to mediate this to each of these receptors suggest that the sensitivity with which high-affinity interaction. The lack of high-affinity binding in responsive cells bind NGF may also depend upon an interaction HSV-x l-infected cells was not due to insufficient expression of between both of these receptors. this construct at the cell surface, since a lo-fold increase in the Measurements of lZSI-NGF binding to ~75~~~~ reveal a single MO1 did not change the Kd (data not shown). Moreover, the low-affinity class of receptors with a Kd = 1O-9 M (Chao et al., uninfected parental cell line, 3T3-trk fibroblasts, as well as con- 1986; Radeke et al., 1987; Rodriquez-Tebar et al., 1990). Sim- trol cells infected with pHSVlac virus displayed only a single ilar measurements of p140pro’01rkindicate that the majority of low-affinity (Kd = 10e9 M) binding site. sites display a low-affinity Kd (Hempstead et al., 199 1; Kaplan From these observations we can propose several hypothetical et al., 199 la; Klein et al., 199 la). A small percentage of higher- models to describe the high-affinity receptor complex. Four of affinity binding sites have been observed for the trk receptor these models are shown in Figure 7. Model A demonstrates a (Klein et al., 199 la). These results have led to several models p75NGFR-p 140 prototrkheterodimer. This model assumes that of high-affinity NGF binding, one including both ~75 and the pl40~~~‘~“~ signals as a monomer. Models B-D assume that trk receptor and another in which p 140prototrk acts independently p 140~ro101rkis activated, like other receptor tyrosine kinases (Ull- as a high-affinity receptor site for NGF (Bothwell, 199 1; Ross, rich and Schlessinger, 1990), as a homodimer interacting with 1991). ~75~~~~. Whether this interaction requires a ~75~~~~ monomer In a further attempt to define the requirements for high-af- (model B), dimer (model C), or multimer (model D) is presently finity NGF binding, we have expressed ~75 cDNAs in defective unclear. In PC12 cells, ~75~~~~ appears to be in excess of the HSV-1 vectors (Geller and Breakefield, 1988). Here we have trk receptor. Furthermore, membrane fusion data has suggested reported the construction of three novel HSV-1 vectors that that the stoichiometric ratio of ~75~~~~ to p140pro101rkfor the encode either p75hNGFRor mutant forms of the receptor. pHSV- reconstitution of high-affinity binding is about 10: 1 (Hempstead hNGFR contains the full-length human ~75 cDNA. The two et al., 1991). A trk dimer-p75 multimer (model D) may be mutant constructs, pHSV-x 1 and pHSV-SA, represent large in- consistent with these observations. tracellular and extracellular deletions, respectively. Direct in- An important next step toward the understanding of the phys- fection of cultured fibroblasts with each of these three HSV-1 iological role of p75NGFR will be to assess the function of this vectors resulted in the transcription and correct processing of receptor in neurons, rather than in heterologous cultured cell receptor RNAs, such that high levels of surface-bound receptor lines. HSV- 1 vectors offer an effective method to introduce cloned protein were detected. These high levels of expression for ~75~~~~ genes in nerve cells (Geller, 1988; Geller and Breakefield, 1988; are required for high-affinity binding. Chiocca et al., 1990; Geller et al., 1991; Federoff et al., 1992). The binding properties of each virally encoded construct were In fact, recombinant HSV-1 vectors provide the only method then assessed by equilibrium binding and Scatchard analysis for direct gene transfer and stable in postmitotic following direct infection into 3T3-trk cells, a fibroblast cell line cells. Here, we have also demonstrated that the HSV-1 vectors that stably overexpress the trk proto-oncogene (> 200,000 sites employed in this study can faithfully deliver high levels of NGF per cell). HSV-hNGFR-infected cells revealed a distinctive two receptors into neurons in primary culture with minimal cyto- site Scatchard plot, consistent with both high- affinity (Kd = 5.11 toxicity. x 10-l’ M) and low-affinity (Kd = 3.99 x 1O-9 M) NGF binding. Infection of cortical neurons in culture with HSV-hNGFR This binding curve parallels that of PC12 cells and sensory (MO1 = 0.25) resulted in high levels of receptor expression as neurons, which are known to possess both classes of receptors detected by both affinity cross-linking and immunofluorescence. (Schecter and Bothwell, 1981; Green and Greene, 1986). The In mock-infected cortical neurons, ~75~~~~ was not detectable, total number of high-affinity sites observed represents only a suggesting that endogenous receptor was either expressed at very small proportion of the total number of NGF binding sites. low levels or not present at this developmental stage in rat However, from our studies it is difficult to assess the absolute cortex. Previous work in feline cortex has demonstrated a tran- number of each of these two NGF binding sites on a per cell sient immunohistochemical staining of p75 in subplate neurons basis because not every cell was infected with vector. At the during brain development (Allendoerfer et al., 1990). However, MO1 used in these experiments, 20-25% of the cells were ex- Yan and Johnson (1988) failed to detect ~75 expression in the pressing the transgene as detected by immunofluorescence. developing rat cortex. Moreover, if endogenous receptor were The Journal of Neuroscience, March 1993, 33) 949

Figure 7. Hypothetical models of the high-affinity NGF receptor complex. Reconstitution of high-affinity NGF binding following infection of 3T3-t&cells with HSV-hNGFR suggests several possible models. ModelA represents a p75-~140 heterodimer. Models B-D depict a trk homodimeric signaling receptor, interacting with either a p75 monomer (model B), dimer (model C), or multimer (model 0). See Discussion.

being expressedat significant levels in theseembryonic cells, it CR, Sehgal A (1986) Gene transfer and molecular cloning of the would have beendetected by affinity cross-linking,an assaythat human NGF receptor. Science 233.518-521.-.__- _-.. does not discriminate the human from the rat receptor. From Chiocca EA, Choi BB, Cai W, DeLuca NA, Schaffer PA, DiFiglia M, Breakefield X0, Martuza RL (1990) Transfer and expression of the these data it would seem that El6 cortex presentsa suitable IucZ gene in rat brain neurons mediated by herpes simplex virus recipient nerve cell population for studiesof p7YGFRfunction mutants. New Biol 2:739-746. in primary culture and in vivo. Chomczynski P, Nicoletta S (1987) Single step method of RNA iso- In summary, we have confirmed by equilibrium binding that lation by acid guanidinium thiocyanate-phenol-chloroform extrac- p 14Opm101*binds to NGF with a single,low affinity in uninfected tion. Anal Biochem 162:156-159. 3T3-trk cells; that p7YGFR participates with p140~~~~~‘*in the Cordon-Card0 C, Tapley P, Jing S, Nanduri V, O’Rourke E, Lambelle F, Kovary K, Klein R, Jones KR, Reichardt LF, Barbacid M (199 1) formation of the high-affinity NCF binding complex, following The trk protein kinase mediates the mitogenic properties of nerve direct infection of 3T3-trk cells with HSV-hNGFR; and that growth factor and neurotrophin-3. Cell 66: 173-l 83. neither the extracellular ligand binding domain nor the cyto- Davies AM (1991) Nerve growth factor synthesis and nerve growth plasmic region of the ~75 receptor when coexpressedwith factor receptor expression in neural development. Int Rev Cytol 128: p140~0101rkin fibroblasts is sufficient to mediate a high-affinity 109-138. Davison AJ, Wilkie NM (1981) Nucleotide secmences of the ioint interaction. We have also shown in this study that our HSV-1 between L and S segments of herpes simplex v&us types 1 and 2. J vectors can rapidly deliver high levelsof functional p75 receptor Gen Virol 55:315-331. in both cultured cell lines and in primary neurons,and further, Davison MJ, Preston VG, McGeoch DJ (1984) Determination of the we have identified E 16 cortical neuronsas a suitable, p75-neg- sequence alteration in the DNA of the herpes simplex virus type 1 ative cell population for future investigations of the physiolog- temperature-sensitive mutant ts K. J Gen Virol 3 1:360-369. ical significance of high-affinity NGF receptors. di Porzio U, Daquet M-C, Glowinski J, Prochiantz A (1980) Effect of striatal cells on in vitro maturation of mesencephalic dopaminergic neurons grown in serum-free conditions. Nature 288:370-373. References Ernfors P, Ibanez CF, Ebendal T, Olson L, Persson H (1990) Molecular Allendoerfer KL, Shelton DL, Shooter EM, Shatz CJ (1990) Nerve cloning and neurotrophic activities of a protein with structural sim- immunoreactivity is transiently associated with ilarities to nerve growth factor: developmental and topographical ex- the subplate neurons of the mammalian cerebral cortex. Proc Nat1 pression in the brain. Proc Nat1 Acad Sci USA 87:5454-5458. Acad Sci USA 87:187-190. Federoff HJ, Geschwind MD, Geller AI, Kess!er JA (1992) Expression Barde Y-A (1989) Trophic factors and neuronal survival. of nerve growth factor in vivo from a defective herpes simplex virus 2:1525-1534. 1 vector prevents effects of axotomy on sympathetic ga&lia. Proc Berkemeier LR, Winslow JW, Kaplan DR, Nikolics K, Goeddel DV, Nat1 Acad Sci USA 89:1636-1640. Rosenthal A (1991) Neurotrophin-5: a novel neurotrophic factor Geller AI (1988) A new method to propagate defective HSV-1 vectors. that activates trk and trkB. Neuron 7:857-866. Nucleic Acids Res 16:5690. Block T, Bothwell M ( 1983) The nerve growth factor receptor on PC1 2 Geller AI, Breakefield X0 (1988) A defective HSV- 1 vector expresses cells: interconversion between two forms with different binding prop- Escherichiu coli @-galactosidase in cultured peripheral neurons. Sci- erties. J Neurochem 40: 1645-l 663. ence 241:1667-1669. Bothwell M (199 1) Keeping track of neurotrophin receptors. Cell 65: Geller AI, During M J, Neve RL ( 199 1) Molecular analysis of neuronal 915-918. physiology by gene transfer into ne&ons with herpes simplex virus Buck CR, Martinez HJ, Black IB, Chao MV (1987) Developmentally vectors. Trends Neurosci 14:428-432. regulated expression of the nerve growth factor receptor gene in the Glass DJ, Nye SH, Hantzopoulos P, Macchi MJ, Squint0 SP, Goldfarb periphery and the brain. Proc Nat1 Acad Sci USA 84:3060-3063. M, Yancopoulos GD (199 1) trkB mediates BDNF/NT-3 dependent Chandler CE, Parsons LM, Hosang M, Shooter EM (1984) A mono- survival and proliferation in fibroblasts lacking the low affinity NGF clonal antibody modulates the interaction of nerve growth factor with receptor. Cell 66:405413. PC12 cells. J Biol Chem 259:6882-6889. Gnahn H, Hefti F, Heumann R, Scwab ME, Thoenen H (1983) NGF- Chao MV (1992) Neurotrophin receptors: a window into neuronal mediated increase in choline acetyltransferase (Chat) in neonatal rat differentiation. Neuron 9:583-593. forebrain: evidence for a physiological role of NGF in the brain? Brain Chao MV, Bothwell MA, Ross AH, Koprowski H, Lanahan AA, Buck Res 285:45-52. 950 Battleman et al. * ~75”~~~~ Defines High-Affinity NGF Binding

Godfrey EW, Shooter EM (1986) Nerve growth factor receptors on roblastoma cells transfected with the NGF receptor gene when treated chick embryo sympathetic ganglion cells: binding characteristics and with NGF. Prog Clin Biol Res 366:227-233. development. J Neurosci 6:2543-2550. McGeoch DJ, Dolan A, Donald S, Brauer DHK (1986) Complete Green SH, Greene LA (1986) A single M, approximately 103,000 lz51- DNA sequence of the short repeat region in the genome of herpes p-nerve growth factor-affinity labeled species represents both the low simplex type 1. Nucleic Acids Res 14: 1727-l 745. and the high affinity forms of the nerve growth factor receptor. J Biol Milbrandt J (1986) Nerve growth factor rapidly induces c-fos mRNA Chem 261:15316-15326. in PC12 rat nheochromocvtoma cells. Proc Nat1 Acad Sci USA 83: Grob P, Berlot CH, Bothwell MA (1983) Affinity labeling and partial 47894793. _ purification of nerve growth factor receptors from rat pheochromo- Miller RH, Hyman RW (1978) Palindrome and palindrome-like se- cvtoma and human melanoma cells. Proc Nat1 Acad Sci USA 80: auences of heroes simnlex virus DNA. Viroloev 87:3a 1. 6819-6823. Mobley WC, Ruttkowski JL, Tennekoon GI, GFmski J, Buchanan K, Hall CV, Jacob GM, Ringold F, Lee J (1983) Expression and regulation Johnston MV (1986) Nerve growth factor increases choline acetyl- of Escherichia coli IacZ aene fusions in mammalian cells. Mol Appl transferase activity in developing basal forebrain neurons. Mol Brain Genet 2:101-109. - Res 1:53-62. Hallbook F, Ibanez C, Persson H (199 1) Evolutionary studies of the Munson P, Rodbard D (1980) LIGAND: a versatile computerized ap- nerve growth factor family reveal a novel member abundantly ex- proach for characterization of ligand binding systems. Anal Biochem pressed in the Xenopus . Neuron 6:845-858. 107:220-239. Hefti F (1986) Nerve growth factor (NGF) promotes survival of septal Oppenheim RW (1989) The neurotrophic theory and naturally oc- cholinergic neurons after fimbrial transection. J Neurosci 6:2155- curring motoneuron death. Trends Neurosci 12:252-255. 2162. Parker BA, Stark CR (1987) Regulation of simian virus 40 transcrip- Hempstead BL, Schleifer LS, Chao MV (1989) Expression of func- tion: analysis of the RNA species present early in infections by virus tional nerve growth factor receptors after gene transfer. Science 243: or viral DNA. J Virol 31:360-369. 373-375. Pleasure S, Reddy UR, Venkatakrishnan G, Roy AK, Chen J, Ross AH, Hemnstead BL, Patil N, Thiel B, Chao MV (1990) Deletion of cy- Trojanowski JQ, Pleasure DE, Lee V (1990) Introduction of nerve toplasmic sequences of the nerve growth factor receptor leads to loss growth factor (NGF) receptors into a medulloblastoma cell line results of hieh affinitv liaand bindinn. J Biol Chem 17:9595-9598. in expression of high- and low-affinity NGF receptors but not NGF- Hempstead BL,-M&tin-Zancab, Kaplan DR, Parada LF, Chao MV mediated differentiation. Proc Nat1 Acad Sci USA 87:8496-8500. (1991) High-affinity NGF binding requires coexpression of the trk Radeke MJ, Misko TP, Hsu C, Herzenberg LA, Shooter EM (1987) proto-oncogene and the low-affinity NGF receptor. Nature 350:678- Gene transfer and molecular cloning of the rat nerve growth factor 683. receptor. Nature 325:593-597. Herrup K, Shooter EM (1975) Properties of the beta-nerve growth Rodriquez-Tebar A, Dechant G, Barde Y-A (1990) Binding of brain factor receptor in development. J Cell Biol 67: 118-l 25. derived neurotrophic factor to the nerve growth factor receptor. Neu- Hohn A, Leibrock J, Bailey K, Barde Y-A (1990) Identification and ron 4:487-492. characterization of a novel member of the nerve growth factor/brain- Rosenthal A, Goeddel DV, Nguyen T, Lewis M, Shih A, Laramee GR, derived neurotrophic family. Nature 344:339-34 1. Nikolics K, Winslow JW (1990) Primary structure and biological Holtzman DM, Li Y, Parada LF, Kinsman S, Chen C-K, Valletta JS, activity of a novel human neurotrophic factor. Neuron 4:767-773. Zhou J, Long JB, Mobley WC (1992) ~140”~ mRNA marks NGF- Ross AH (199 1) Identification of tyrosine kinase Trk as a nerve arowth responsive forebrain neurons: evidence that trk gene expression is factor receptor. Cell Regul 2:6851690. induced by NGF. Neuron 9:465478. Ross AH. Grob P. Bothwell M. Elder DE. Ernst CS. Marano N. Christ Ibanez CF, Ebendal T, Barbany G, Murray RJ, Blundell TL, Persson BFD, Slemp CC, Herlyn M, Atkinson B; Koprowski H (1984) Char- H (1992) Disruption of the low affinity receptor-binding site in NGF acterization of nerve growth factor receptor in neural crest tumors allows neuronal survival and differentiation by binding to the trk gene using monoclonal antibodies. Proc Nat1 Acad Sci USA 8 1:668 l-6685. product. Cell 69:329-34 1. Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a lab- Johnson D, Lanahan A, Buck CR, Sehgal A, Morgan C, Mercer E, oratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Lab- Bothwell M, Chao MV (1986) Expression and structure of the hu- oratory. man NGF receutor. Cell 47:545-554. Scatchard G (1949) The attraction of proteins for small molecules and Kaplan DR, Hempstead BL, Martin-Zanca D, Chao MV, Parada LF ions. Ann NY Acad Sci 52:660. (i 99 1a) The trk proto-oncogene product: a signal transducing recep- Schechter AL, Bothwell MA (198 1) Nerve growth factor receptors on tor for nerve growth factor. Science 252:554-557. PC 12 cells: evidence for two receptor classes with different cytoskel- Kaplan DR, Martin-Zanca D, Parada LF (1991 b) Tyrosine phos- eta1 associations. Cell 24:867-874. phorylation and tyrosine kinase activity of trk proto-oncogene product Schechterson LC, Bothwell MA (1992) Novel roles for neurotrophins induced bv NGF. Nature 350: 158-l 60. are suggested by BDNF and NT-3 mRNA expression in developing Klein R, Jing S, Nanduri V, O’Rourke E, Barbacid M (199 la) The neurons. Neuron 9:867-874. trk proto-oncogene encodes a receptor for nerve growth factor. Cell Soppet D, Escandon E, Maragos J, Middlemas DS, Reid SW, Blair J, 65:189-197. Burton LE, Stanton BS, Kaplan DR, Hunter T, Nikolics K (1991) Klein R, Nanduri V, Jing S, Lambelle F, Tapley P, Bryant S, Cordon- The neurotrophic factors brain-derived neurotrophic factor and Cardo C, Jones KR, Reichardt LF, Barbacid M (1991 b) The trkB neurotrophin-3 are ligands for the trkB tyrosine kinase receptor. Cell tyrosine protein kinase is a receptor for brain-derived neurotrophic 65:895-903. factor and neurotrovhin-3. Cell 66:395403. Spaete RR, Frenkel N (1982) The herpes simplex virus amplicon: a Lambelle F, Klein R, Barbacid M (199 1) trkC, a new member of the new eukaryotic defective-virus cloning-amplifying vector. Cell 30: trk family of tyrosine kinases, is a receptor for neurotrophin-3. Cell 295-304. 661967-979. Squint0 SP, Stitt TN, Aldrich TH, Davis S, Bianco SM, Radziewski C, Landreth GE, Shooter EM (1980) Nerve growth factor receptors on Glass DJ, Masiakowski P, Furth ME, Valenzuala DM, DiStephano PC1 2 cells: ligand-induced conversion from low- to high-affinity states. PS, Yancopoulos GD (199 1) trkB encodes a functional receptor for Proc Nat1 Acad Sci USA 77:4751-4755. brain-derived neurotrophic factor and neurotrophin-3 but not for Leibrock J, Lottspeich AH, Hofer M, Hengerer B, Masiakowski P, Theo- nerve growth factor. Cell 65:885-893. nen H, Barde Y-A ( 1989) Molecular cloning and expression of brain- Sutter A, Riopelle RJ, Harris-Warrick RM, Shooter EM (1979) Nerve derived neurotrophic factor. Nature 34 1: 149-152. growth factor receptors. Characterization of two distinct classes of Levi-Montalcini R (1987) The nerve growth factor 35 years later. binding sites on chicken embryo sensory ganglia cells. J Biol Chem Science 237: 1154-l 162. 25415972-5982. Maisonpierre PC, Belluscio L, Squint0 S, Ip NY, Furth ME, Lindsay Thoenen H, Barde Y-A (1980) Physiology of nerve growth factor. RM. Yanconoulos GD (1990) Neurotrophin-3: a neurotrophic fac- Physiol Rev 60: 1284-1334. tor related to NGF and BDNF. Science 247: 1446-145 1. Ullrich A, Schlessinger J (1990) Signal transduction by receptors with Matsushima H, Bogenmann E (199 1) Terminal differentiation in neu- tyrosine kinase activity. Cell 6 1:203-2 12. The Journal of Neuroscience, March 1993, 13(3) 951

Welcher AA, Bitler CM, Radeke MJ, Shooter EM (199 1) Nerve growth Yan H, Chao MV (1991) Disruption of cysteine-rich repeats of the factor binding domain of the nerve growth factor receptor. Proc Nat1 p75 nerve growth factor receptor leads to loss of ligand binding. J Acad Sci USA 88:159-163. Biol Chem 266:12099-12104. Weskamp G, Reichardt LF (1991) Evidence that biological activity Yan Q, Johnson EM (1988) An immunohistochemical study of the ofNGF is mediated through a novel subclass ofhigh affinity receptors. nerve growth factor receptor in developing rats. J Neurosci 8:348 l- Neuron 61649-663. 3498.