THE ~VURNALof BlOLoCICALCHEMISTRY Vol. 256. Nn. ti. hue of March 25. pp. 22%-2292, 1980 Prmtrd in USA. Effect of Guanosine Diphosphate on Microtubule Assemblyand Stability* (Received for publication, April 23, 1979, and in revised form, August 10, 1979) Larry Jameson andMichael Caplow From the Department of Biochemistry, University of North Carolina, ChapelHill, North Carolina 27514 Polymerization of microtubular protein is promoted formation, only guanine nucleotides are bound to tubulinwith by GTP but not GDP. GDP is shown to inhibit both the high affinity (2, 3). The tubulin dimer has two binding sites rate and extent of GTP-induced microtubule assembly. for guanine nucleotides. One site is nonexchangeable (N-site)’ The critical protein concentration for assembly was and releases GTP only upon denaturation. The other site(E- increased from 0.3 mg/ml with 200 JLM GTP to 0.8 mg/ site) readily exchanges bound nucleotide for GDP or GTPin ml with 200 JLM GTP and 600 JLM GDP. The relative the medium (3). The nonguanine nucleoside triphosphates binding constants for GDP and GTP to the tubulin have been proposed to induce tubulin polymerization indi- exchangeable binding site (E-site) are used to quanti- rectly by phosphorylating GDP througha nucleoside diphos- tatively analyze the mechanism by which GDP raises phokinase reaction (3-5). GTP promotes polymerization by the critical protein concentration and thereby reduces interacting directly with the tubulin E-site(3-5). the extent of microtubule assembly. Evidence that guaninenucleotides also interact with GTP hydrolysis accompanies microtubule polymerization microtubular protein at a previously unknown, low (6-8). However, it has been difficult to determine whether affinity binding site is supported by the following ob- GTP hydrolysis is required for microtubule assembly or dis- servations. assembly, if either. Isolated microtubules contain only GDP 1. At a constant ratioof GDP/GTP, a greater degree at the E-site (3, 4, 9, lo), indicating that phosphate bond of inhibition of the extentof microtubule assemblywas hydrolysis occurs after microtubule assembly but before dis- observed at higher concentrations of total nucleotide. assembly. Nevertheless, analogs of GTP which are resistant This result cannot be accounted for by binding at only to hydrolysis at the p-y phosphate bond (e.g. Gpp(NH)p) the E-site. have been shownto promote microtubuleassembly (11). Since 2. Inhibition of microtubule assembly by a high con- the microtubulesformed using these analogs exhibit unusual centration of GDP (3 mM) is not reversed by increasing stability in the presence of concentrations of calcium which the GTP concentration to 7 m. Based upon the disso- completely depolymerize microtubules formed with GTP, it ciation constants forGDP and GTP at theE-site, tubulin has been suggested that hydrolysis of the terminal phosphate would be largely saturated with GTP under these con- of GTP may be necessary before microtubule disassembly can ditions and should polymerize in the absence of addi- occur (4). However, studies using another analog of GTP tional nucleotide interactions. which is resistant to phosphate bond hydrolysis at the a-p 3. GTP specifically inhibits microtubule assembly at position but not thep-y position (Gp(CH2)pp)give somewhat concentrations greater than3 mM. Inhibition of assem- different results. Using a quantitative colchicine binding assay bly by high concentrations (7 m~)of ATP and CTP is for tubulin polymerization, Sandoval et al. (12) demonstrated shown to involve a differentmechanism. that Gp(CH2)pp promotespolymerization of 98% of the avail- 4. Microtubules are depolymerized by high concen- able tubulin into microtubules, while only 70% of the tubulin trations of GTP (>5 mM) or GDP (>2 mM). The ratefor this disassembly is greater than thatobserved during was polymerized in the presence of GTP. Although the ter- microtubule turnover following inhibition of polymer- minal phosphate of Gp(CH8)pp was hydrolyzed,the rnicrotu- ization, suggestingthat high Concentrations of guanine bules formed with thisanalog were still resistant todepolym- nucleotides destabilize microtubules by interacting at erization by calcium. This observation was supported by ad- a low affinity site. ditional studieswhich showed that microtubulescould also be formed using the nonhydrolyzable analog of GDP, Gp(CHz)p, in either thepresence or absenceof calcium (13).These results suggest that hydrolysis of the a-P phosphate bond may be Microtubules are assembled from the dimeric protein sub- important for microtubule disassembly. unit, tubulin. The mechanism by which the assembly and The role of the GDPin microtubule assembly and disassem- disassembly of microtubules are regulatedin vivo is unknown. bly must be further elucidated before definitive conclusions However, the characterization of conditions which favor as- can be drawn concerning the function of p-y or a-p phosphate semblyor disassembly in vitro has progressed rapidly in bond hydrolysis. A number of laboratories have reported that recent years (for review, see Ref. 1). In particular, studies of GDPdoes not promote tubulin polymerization (4, 6, 13). tubulin-nucleotide interactions have revealed that microtu- bule assembly requires nucleoside triphosphates. Whilea ’ The abbreviations used are: N-site, nonexchangeable binding site; variety of nucleoside triphosphates can induce microtubule E-site, exchangeable binding site; Gpp(NH)p, guanyl-5”yl imidiphos- phate; Gp(CH2)pp,guanyl-5’-yl methylene diphosphonate; Gp(CHn)p, * This work was supported by Grant DE03246 from the National guanosine 5”methylene diphosphonate; HB, reassembly buffer (0.1 Institute for Dental Research. The costs of publication of this article M 2-(N-morpholino)ethanesulfonicacid buffer, 0.5 m~ MgClz, 1.0 mM were defrayed in part by the payment of page charges. This article EGTA, pH 6.8); MAPS, microtubule-associated proteins; PEI, poly- musttherefore be hereby marked “advertisement” in accordance ethyleneimine; EGTA,ethylene glycol bis(8-aminoethyl ether)N,- with 18 U.S.C. Section 1734 solely to indicate this fact. N,N‘,N’-tetraacetic acid. 2284 This is an Open Access article under the CC BY license. M icro tubu le Assembly and Disassembly and AssemblyMicrotubule 2285 Furthermore, GDP has been shown to inhibit GTP-induced by thin layer chromatography on PEI-cellulose as described previ- microtubuleassembly (4, 6, 8). Carlier and Pantaloni (14) ously (19). GTP hydrolysis was determined by the decrease in ["PIGTP relative to recovered ['HIGTP. recently studied polymerization of phosphocellulose-purified Measurement of Transphosphorylase Activity-Nucleoside di- tubulin in the presence of 5 mM Mg2+ and 3.4 M glycerol and phosphokinase (transphosphorylase) activity was determined by fol- found that in contrast to earlier reports, the tubulin. GDP lowing the conversion of ['HIGDP to [:'H]GTP in the presence of complex could participatein microtubule elongation after the added ATP. Shortly before use, ["HIGDP was purified by thin layer nucleation stepwas carried out usingGTP. chromatography on PEI-cellulose in order to remove radioactive The purpose of this report is to clarify the role of GDP in contaminants. Reaction mixtures containing 15 pI of ATP (20 mM or 200mM), 35 of purified vH]GDP (8500 cpm/pl), and 400 1.11 of microtubuleassembly disassembly. find the pl and We that tubulin (31 p~)were incubated at either 4OC or 37°C. The reactions tubulin.GDP complex does not participate in either the nu- were stopped at different times by centrifuging 5O-pl aliquots through cleation or the elongation of microtubules. GDP inhibition of 1-ml columns of centrifuge-packed G-25 to separate the protein from the rate and extent of GTP-induced microtubule assembly is exogenous nucleotide (ATP) (20). In separate experiments, it was analyzed in terms of a mechanism in whichthe tubulin-GDP found that only guanine nucleotides remain protein-bound under complex is not competent to polymerize. Evidence for a low these conditions, and that ATP and ADP are entirely removed by this procedure. The protein was eluted into 2% perchloric acid and affinity binding site for guanine nucleotides is presented. the released nucleotides isolated and separated by thin layer chro- matography on PEI-cellulose (19). Transphosphorylase activity was EXPERIMENTALPROCEDURES determined from the relative amounts of"H in the resolved spots Methods containing GTP and GDP. The recoveries of [ 'HIGDP and [ 'H1GTP are identical in the purification procedure (determined in separate Preparation of Microtubular Protein-Microtubular protein was experiments), so that this method allows determination of the relative prepared by a modification of the Shelanski (15) assembly-disassem- radioactivity in the nucleoside di- andtriphosphates in reaction bly procedure. Fresh pig brains were homogenized with ice-cold mixtures. reassembly buffer (I ml/g of brain) (0.1 M 2-(N-morpholino)- ethanesulfonic acid buffer, 0.5 mM MgClz,1.0 mM EGTA, pH 6.8) Materials containing 0.1 mM ATP in a Waring Blendor for 30 s. The supernatant Guanyl-5"yl methylene diphosphonate (Gp(CHz)pp, which had obtained from centrifugation (30,000 rpm, 75 min, 4°C) in a Beckman been purifiedby high pressure liquid chromatography, was purchased 30 rotor was diluted with an equal volume of glycerol-reassembly from ICN. ["]GTP (specific activity 10.6 Ci/mmol) and [:'H]GDP buffer (60.2 ml of glycerol: 24 ml of reassembly buffer) containing 1 (specific activity, 8.91 Ci/mmol) and