Proc. Nat. Acad. Sci. USA Vol. 71, No. 5, pp. 2131-2135, May 1974 Interaction of Nerve Growth Factor with the Mouse-Brain Neurotubule Protein(s) (tubulin/divalent cations/colchicine/GTP) PIETRO CALISSANO AND COSTANTINO COZZARI Laboratory of Cell Biology, Via Romagnosi 18A, Rome, Italy Communicated by R. Levi-Montalcini, January 21, 1974 ABSTRACT Addition of nerve growth factor to a 105,000 Purification of Tubulin. The microtubule protein was puri- X g supernatant of mouse brain induces the formation of fied by the method of Shelanski et al. (8) from adult mouse a precipitate whose main constituent is the microtubule protein(s) (tubulin). The binding of nerve growth factor to brain with a minor modification. The composition of the re- purified tubulin is not inhibited by colchicine and does not assembly buffer was 10 mM KH2PO4-Na2HPO4, pH 6.5, in- appear to depend on the presence of GTP or Mg++. GTP, stead of 100 mM MES; 1 mM EGTA, 1 tnM GTP, and 0.5 however, and divalent cations, exert a marked effect on the mM MgCl2 were also present to form the reassembly buffer increased turbidity induced by interaction of nerve growth Tubulin was 80- factor with tubulin. These findings are tentatively in- as described by Weisenberg (6). generally terpreted with the hypothesis that binding of the factor to 90% pure, as judged by densitometric scanning after sodium tubulin and the induced aggregation is a sequential two- dodecyl sulfate gel electrophoresis (see also Fig. ld). For step process; the latter but not the former would be in- binding studies, aliquots of tubulin in 8 M glycerol were di- fluenced by GTP or divalent cations. luted to 4 M with reassembly buffer, incubated for 20 min The most striking effect of nerve growth factor (NGF) is at 370, and centrifuged at 100,000 X g for 60 min. The pellet stimulation of rapid neurite outgrowth (1). This effect is was resuspended in the reassembly buffer, except where other- limited to embryonic sensory nerve cells during a restricted wise stated, and kept at 20 for at least 30 min to ensure de- period of their development and sympathetic nerve cells polymerization. Before use for binding studies this preparation during all developmental stages (1, 2). One of the earliest was centrifuged at 3000 rpm for 30 min. effects elicited by NGF is massive production of neurofila- 125I-Labeled NGF and [3H]Colchicine Binding to Tubulin. ments and neurotubules, filling the cytoplasm. This response was measured the filter method is detectable as early as 2 hr (3, 4) after the beginning of incu- [3H]Colchicine binding by (6), with an incubation at 370 for 30 min. '25I-Labeled NGF bation with NGF. A quantitative analysis of this effect has was measured of been binding by incubating aliquots purified reported (5). tubulin (generally 100-150 fig) for 30 min at 370 in 0.25 ml of The neurotubules or microtubules are formed by the as- also 0.1 M NaCl and 2 a the reassembly buffer containing mg/ sembly of precursor, dimeric protein (molecular weight ml of a mixture of order to reduce ad- 110,000) called microtubule protein or tubulin (6-8). The immunoglobulins (in sorbtion of the NGF to the test tube) plus constant amounts dimer appears to be composed of two related but not identical of 125J-labeled NGF and various concentration of unlabeled subunits (9) and is characterized by its unique ability to bind colchicine, so that it is also referred to as colchicine- NGF. After incubation, the tubes were mixed with 0.5 mM vinblastine, further incubated for 15 min, and then cen- binding protein (10). Tubulin is ubiquitous within the cell for 8 min in a Beckman 152 at room tem- cytoplasm in a soluble and in a membrane-bound form (11, trifuged microfuge perature. The counts in the pellet were taken as a measure of 12). the extent of protein bound. Vinblastine facilitates precipita- We report here on a specific, high-affinity binding of NGF tion of the tubulin-NGF thus the need of to this protein. This finding appears to be pertinent to the complex, avoiding mechanism of action of NGF as well as to its still unknown high-speed centrifugation without interfering with the bind- receptor. ing, once the complex is formed. Controls without tubulin, at every NGF concentration, were run in the same experiment in MATERIALS AND METHODS order to subtract the nonspecific contribution due to spon- Purification of NGF. NGF was prepared by the method of taneous precipitation or adsorption of NGF to the test tube, Bocchini and Angeletti (13), and its molar concentration was which never exceeded 5-8% of the total NGF bound. based on the assumption of a molecular weight of 28,000 (14). 125I-Labeled NGF, a generous gift of Dr. Roberto Revoltella, Light Scattering Measurements. The interaction between NGF had a specific activity of 10 cpm/mg of protein when counted and tubulin was followed by measurement of the light scat- on a Wallac gamma counter (LKB) GTL 300-500 equipped tering of tubulin at 400 nm after addition of NGF. Generally, with a NaI crystal. The labeled NGF retained full biological to a solution of tubulin (50-100 Mg) in the reassembly buffer, activity, as determined by its effect in vitro on chick-embryo aliquots of NGF were added in a final volume of 0.25 ml. sensory ganglia. The solution was rapidly stirred, and the increase in absor- bancy (A) was monitored with an MQ III Zeiss spectro- Abbreviations: NGF, nerve growth factor; MES, 2-(N-morpho- photometer. The spontaneous aggregation or polymerization lino)ethanesulfonic acid; EGTA, ethylene glycol bis(G-amino- of tubulin (8, 15) in the absence of NGF that occurs in the ethyl ether)-N,N'-tetraacetic acid. reassembly buffer never exceeded 5-7% of that induced by 2131 Downloaded by guest on September 29, 2021 2132 Cell Biology: Calissano and Cozzari Proc. Nat. Acad. Sci. USA 71 (1974) ,xi CD a 20. 'S 0 m LL 2 10 1 104 f% 20%If% 3u-In WP-' 70--:. NGF (mg) FIG. 2. Binding of NGF to tubulin in the presence and absence of colchicine. The assay was performed with 60 jg of purified tubulin plus various amounts of NGF (from 100 to 1.5 Mug) and other substances (see Methods) in a final volume of 0.25 ml. At the end of incubation (30 min at 370), complex was pre- cipitated with 0.5 mM vinblastine. The concentrations of NGF on the abscissa refer to free NGF at the end of incubation, which + cm is given by total NGF minus the fraction of the protein bound at the end of the experiment. (0) NGF alone; (0) plus 0.1 mM FIG. 1. Sodium dodecyl sulfate electrophoresis after NGF or colchicine. vinblastine treatment of brain supernatants. Two mouse brains were homogenized with a 3:1 v/w of 10mM MES (pH 6.5), 0.5 mM MgCl2, 1.0 mM EGTA, and 1.0 mM and X g supernatant of mouse brain, an almost instantaneous tur- GTP, centrifuged bidity of the protein solution occurs. After for 30 at 105,000 X g for 90 min. The supernatant was divided into standing min 0.25-ml aliquots (0.8 mg of total protein), and 0.04 ml of buffer at 20, the solution is centrifuged at 105,000 X g for 30 min and containing 140 Mg of NGF (b) or 0.03 ml of 10 mM vinblastine (c) the resulting pellets, dissolved in the sodium dodecyl sulfate was added. After standing for 30 min at 20, the mixtures were buffer for electrophoresis (see Methods), shows that NGF is centrifuged for 30 min at 105,000 X g. The pellets of NGF- and precipitated mainly with one protein component among all vinblastine-treated samples were resuspended in buffer and cen- proteins present (Fig. la and b). Thus, the band indicated trifuged again at 105,000 X g for 30 min. After centrifugation the by the arrow in Fig; la, which accounts for about 8-10% of sediments were dissolved in 0.1 ml of 1% sodium dodecyl sulfate- the total proteins in the gel, is increased to 40-50% after mercaptoethanol (see Methods). After centrifugation and removal precipitation with NGF, while most of the other proteins are of the pellet, 0.25 ml of the vinblastine-treated supernatant was left in the supernatant. This band has an apparent molecular dialyzed overnight and 140 Mg of NGF was subsequently added. weight of 52,000-53,000 when compared to standard proteins The sample (e) was then treated as described for b and c. (a) 105,000 X g total supernatant; (b) 0.025 ml of the pellet after of known molecular weight (cytochrome c, chymotrypsino- NGF treatment; (c) 0.025 ml of the pellet after vinblastine pre- gen, bovine-serum albumin, and ovalbumin), and shows the cipitation; (d) 50 Mg of purified tubulin (8); (e) 0.025 ml of the same mobility of a purified preparation of tubulin (Fig. ld) pellet after vinblastine treatment, dialysis of supernatant, ad- (see Methods). When the two bands (NGF-precipitated pro- dition of NGF, and recentrifugation. The unlabeled arrows indi- tein and purified tubulin) are cut out and analyzed, they cate the tubulin peak. show an overimposable aminoacid composition. The vin- blastine-induced precipitate (Fig. ic) of an identical 105,000 the lowest amount of NGF tested. The contribution of this X g supernatant is very similar to that obtained with NGF. scattering was subtracted from the induced effect of NGF.