ADP-Ribosylation of Microtubule Proteins As Catalyzed by Cholera Toxin

The EMBO Journal, Vol. 1. No. 2, pp. 181-186, 1982

ADP-Ribosylation of microtubule proteins as catalyzed by cholera toxin

Y.Amir-Zaltsman1, E.Ezral, T.Scherson2, A.Zutra2, U.Z. Results Liftauer2, and Y.Salomon1 Rat brain microtubule proteins were obtained by two Departments of 'Hormone Research and 2Neurobiology, The Weizmann In- assembly-disassembly cycles. The fraction obtained contains stitute of Science, Rehovot 76100, Israel both tubulin and MAPs. Incubation of this fraction with Communicated by U.Z.Littauer [32P]NAD and cholera toxin and subsequent polyacrylamide Received on 28 December 1981 gel electrophoresis under denaturing conditions resulted in the labeling of both the a and (3 tubulin subunits as well as MAP1 Incubation of purified rat brain tubulin with cholera toxin and MAP2 (Figure 1). Further purification of the micro- and radiolabeled [32p] or [8-3H]NAD results in the labeling of tubule proteins by phosphocellulose chromatography both a and (3 subunits as revealed on sodium dodecyl sulfate separated the tubulin from the MAPs. Incubtion of the phos- polyacrylamide gel electrophoresis (SDS-PAGE). Treatment phocellulose-purified tubulin fraction with cholera toxin and of these protein bands with snake venom phosphodiesterase labeled NAD causes considerable labeling of the tubulin resulted in quantitative release of labeled 5'-AMP, respec- subunits (Figure 1, lane 6). It was also noted that the ( sub- tively labeled with the corresponding isotope. Two- unit was labeled more heavily than the a subunit. The posi- dimensional separation by isoelectric focusing and SDS- tion of the radiolabeled a and (3 subunits correlated well with PAGE of labeled and native tubulin revealed that labeling oc- the position of the Coomassie brilliant blue stained bands of curs at least in four different isotubulins. The isoelectric point purified tubulin (lane 8). In addition some radioactivity was of the labeled isotubulins was slightly lower than that of incorporated into high mol. wt. protein bands, which had a native purified tubulin. This shift in mobility is probably due mobility identical to that of marker rat brain MAP1 and to additional negative charges involved with the incorpora- MAP2. It should be noted that the high mol. wt. MAPs are tion of ADP-ribosyl residues into the tubulin subunits. SDS- present at very low concentrations in the purified tubulin PAGE of peptides derived from [32P]ADP-ribosylated a and fraction and were not detectable by Coomassie brilliant blue (3 tubulin subunits by Staphylococcus aureus protease staining yet they appear to be labeled by this reaction. In the cleavage showed a peptide pattern identical with that of absence of cholera toxin no labeling of any of these proteins native tubulin. Microtubule-associated proteins (MAP1 and was observed (Figure 1, lane 7). MAP2) of high molecular weight were also shown to undergo To examine whether additional MAPs are labeled in this ADP-ribosylation. Incubation of permeated rat neuro- reaction, we incubated the MAP fraction, purified by phos- blastoma cells in the presence of [32P]NAD and cholera toxin phocellulose column chromatography, with [32P]NAD and results in the labeling of only a few cell proteins of which cholera toxin. This fraction contains the tau factors, the high tubulin is one of the major substrates. mol. wt. polypeptides (MAPI and MAP2) as well as other Key words: tubulin/MAPs/cholera toxin/ADP-ribosylation proteins (Weingarten, 1975). Figure 1 (lane 2) shows considerable labeling of MAP1 and MAP2, while tau factors 1-4, were only weakly labeled, thus indicating the selective Introduction nature of the cholera toxin-mediated reaction. In the absence ADP-ribosylation has recently emerged as a potentially im- of cholera toxin, no labeling was observed (Figure 1, lanes 3 portant mechanism for controlling the activity of several pro- and 5, respectively). In a control experiment (Figure 1, lane 1) teins (Hayaishi and Veda, 1977; Hinz et al., 1978; Pap- cholera toxin was incubated with [32P]NAD in the absence of penheimer, 1977). Some bacterial toxins have been shown to any other acceptor proteins. Under these conditions, i.e., 807o catalyze NAD-dependent ADP-ribosylation of key regulatory sodium dodecyl sulfate polyacrylamide gel electrophoresis proteins. ADP-ribosylation of elongation factor 2 diphtheria (SDS-PAGE), labeled cholera toxin runs out of the slab gel. toxin (Pappenheimer, 1977) leads to inhibition of protein syn- Consequently one could conclude that the labeled bands seen thesis. Cholera toxin catalyzes ADP-ribosylation of the GTP- in lanes 2, 4, and 6 represent brain proteins. binding protein of the hormone-sensitive adenylate cyclase To permit labeling at high specific radioactivity, experi- system, thereby leading to more persistent cyclase stimulation ments were carried out at micromolar concentrations of (Cassel and Pfueffer, 1978; Johnson et al., 1978; Gill and radiolabeled NAD. Since the Km of cholera toxin for NAD is Meren, 1978). The common denominator of these two accep- high (3-4 mM; Mekalanos, 1979) the fraction of labeled tor proteins is that they both specifically bind and cleave GTP tubulin obtained was minute, and did not exceed 0.1 7o. It and that ADP-ribosylation inhibits their activity. It was, was therefore imperative to identify the labeled proteins by therefore, of interest to test whether tubulin, which apparent- independent methods. We first subjected the presumptive ly is one of the most abundant GTP-binding proteins in labeled tubulin bands to isoelectric focusing followed by eukaryotic cells can serve as acceptor for ADP-ribosylation SDS-PAGE (Figure 2). Protein staining of the purified catalyzed by cholera toxin. tubulin following the two-dimensional electrophoresis show- In this study we demonstrate that rat brain tubulin, as well ed four spots (Figure 2A). Autoradiography of this slab gel as high molecular weight (mol. wt.) microtubule-associated (Figure 2B) revealed a similar picture with respect to the proteins (MAPI and MAP2), serve as substrates for cholera separation of the a and (3 tubulin subunits. Superposition of toxin NAD-dependent ADP-ribosylation. the autoradiogram (Figure 2B) on the Coomassie brilliant

2 323456~~~74 5 6 13

(t) IEF .-...... 2 -MAP l 2 -MAP 2 (A) Protein Stain

T ubulIn ''-' :.wr kG .:'."V.

1 - 2- taou 3- TUBULIN - a - aw 4 - -R- Cholera 1...... 1 Toxin

* ilcm (B) Autoradiogrom £Q C HOLE RA T OX IN" (+) (+) (-) (+) (-) (+) (-) PROTEIN STAIN -.-'----Autor a diogram .---. -

Fig. 1. SDS-PAGE of tubulin and MAPs incubated with [32P]NAD and *7Tubulin cholera toxin. The reaction mixtures contained [32P]NAD 16.3 Ci/mmol and the protein to be labeled with or without cholera toxin as indicated. l,g10 of labeled protein fraction were loaded in each lane, and the resulting autoradiographic pattern is shown. Lane 1, cholera toxin alone with [32P]NAD; lanes 2, 3, MAPs after separation from tubulin by phosphocellulose chromatography incubated with or without cholera toxin; Cholera lanes 4, 5, assembled tubulin (second cycle) incubated with or without ToXin cholera toxin; lanes 6, 7, phosphocellulose-purified tubulin incubated with or without cholera toxin; lane 8 is a Coomassie brilliant blue staining of lane 5. blue stained gel (Figure 2A) showed that the mobility in the second dimension of the various isotubulins on SDS-PAGE was identical. In contrast, the various spots on the autoradiogram were slightly shifted in the acid direction with an Fig. 2. Two-dimensional gel electrophoresis of purified tubulin labeled average ApI of 0.18 for both the a and ,3 subunits. with [32P]NAD. Labeling of purified tubulin was performed as described in Cholera toxin is also labeled by [32P]ADP-ribosylation. Materials and methods using cholera toxin and [32P]NAD 45 Ci/mmol. The labeled toxin well from the tubulin subunits on Labeled purified tubulin (13 Ag protein) was subjected to isoelectric focus- separates ing. The resulting isoelectric focusing gels were separated (second dimen- two dimensional gel electrophoresis as seen in lane 2, Figure sion) by SDS-PAGE (7.5-15% polyacrylamide gradient, 20 cm long). 2B. The spot at the center of Figure 2A (indicated by an ar- Protein staining of the gel (part A) and autoradiogram (part B) are shown. row) originates in the cholera toxin preparation but is not Markers separated on SDS-PAGE only are: lane 1 (A,B) unlabeled tubulin labeled by incubation with radiolabeled NAD. mixed with cholera toxin which has been previously incubated with [32PJNAD under standard labeling conditions but with no tubulin; lane 2 Another approach to verify the identity of the labeled (A,B) contains labeled tubulin. The dashed markings on parts A and B in- tubulin bands was to examine their peptide distribution on dicate the location of the corresponding isotubulin spots as found on the SDS-PAGE following limited digestion with Staphylococcus autoradiogram (part B) and on the protein stained gel (part A) respectively. aureus protease. As shown in Figure 3, proteolytic digestion The arrow (part A) indicates an unlabeled peptide originating from the of a (lanes 2, 3) and (3 tubulin (lanes 5, 6) with S. aureus pro- cholera toxin preparation. tease resulted in different peptide distribution for the two subunits on SDS-PAGE (15 -20o). Staining with It was assumed that if ADP-ribosylation is involved, then a Coomassie brilliant blue (lanes 2, 5) and autoradiography transfer of the adenine moiety of NAD to tubulin should be (lanes 3, 6) indicated that the major cleavage peptides of a observed. We therefore tested the action of cholera toxin on and ( tubulin incorporated label following incubation of purified tubulin using [8-3H]NAD (Figure 4) and we were able purified tubulin with [32P]NAD and cholera toxin. Excised to show that tubulin is indeed labeled in this reaction and that bands of a and (3 tubulin which were not digested with pro- the position of the labeled protein corresponded upon gel tease were loaded in lanes 4 and 7 respectively. electrophoresis to that of unlabeled tubulin (a and (3 tubulin We next examined the nature of the tubulin modification. bands were clearly seen after staining with Coomassie 182 ADP-Ribosylation of microtubule proteins

1 2 2 3 4 5 6 7 M.W.

( x 0o3) -a - V------A- i-200 00 - 9 2.5 69 -a+, F. 4F - 46 ;3"eA:@ W

- 30 .-I , ..:

1~: -i 14.3 (

7- ...N

__ -~~~~~~~j |iProtease: (- )

T U BIUL I N- i'W` Protein Sta in

Autorcdiog rar-ot

PROTEIN 3H Fig. 3. S. aureus protease cleavage peptides of [32P]ADP-ribosylated STAIN purified tubulin. Tubulin (phosphocellulose fraction) was incubated with labeled [32P]NAD 35 Ci/mmol or unlabeled NAD and cholera toxin as described in Materials and methods. The ADP-ribosylated protein was subjected to 8% SDS-PAGE: 5 ug protein/slot for the radiolabeled and 40 yg Hg. of tubulin labeled with [32P]NAD and cholera toxin. protein/slot for the unlabeled purified tubulin. The stained gels were dried 4. SDS-PAGE Tubulin (phosphocellulose fraction) was labeled with [8-3H]NAD (41 Ci/ and the corresponding a and bands were excised, prepared, proteolytic- mmd) in the presence of cholera toxin. 10 jg of labeled tubulin were load- ally digested by S. aureus protease, and subjected to gel electrophoresis as ed in each lane. The 8%7o SDS-PAGE was impregnated with sodium described in Materials and methods. The resulting peptide maps of a (laes as described in Materials and methods. The Coomassie brilliant 2- 4) and (lanes 5-7) subunits of purified tubulin are shown. salicylate blue staining and the autoradiogram are shown. brilliant blue but this pattern diffused following further treat- is ment with salicylate). It should be noted that the [32P]NAD Table I. Distribution of radioactive products released from ADP-ribo- labeled at the a position adjacent to the 5'-carbon of the sylated tubulin after treatment with snake venom phosphodiesterase adenosine moiety. It therefore appears that the radioactively labeled derivatized tubulin contains at least the adenosine and NAD labeling Snake venom Origin 5'-AMP Total the a phosphate of NAD. phospho- (c.p.m. 07o)a c.p.m. In other experiments (Table I), we were able to show that diesterase loaded [3H] or [32p]5 '-AMP are released from 3H- or 32P-labeled tubulin by treatment with snake venom phosphodiesterase. [32P]NAD - 99 < 1 370 < 1 99 2911 Purified 3H- and 32P-labeled tubulin was prepared by incuba- + tion with cholera toxin and [8-3H]NAD or [32P]NAD, respec- [8-3H1INAD - 85 9 494 tively. The labeled protein preparations were subjected to + < I 100 1955 SDS-PAGE on 8% polyacrylamide and the radioactive bands were then excised. The protein was extracted and precipitated a Spots corresponding with NAD as well as the intervening spaces between with 10qo trichloroacetic acid. The resulting pellets of the spots were also extracted and counted but contained no significant radiolabeled tubulin preparations were incubated with snake activity. venom phosphodiesterase for 4.5 h. Reaction products were Purified tubulin (100 ug) was labeled with [32P]NAD 35 Ci/mmol or and on 8% identified by t.l.c. on polyethyleneimine cellulose. Table I [8-3H]NAD 41 Ci/mmol and cholera toxin, electrophoresed SDS-PAGE (10 zg/lane) as described in Materials and methods. Bands shows that quantitative release of [32p] or [3H]5' -AMP from from three lanes were pooled for extraction and digestion by phospho- the respective labeled tubulin bands depended on the presence diesterase as described in Materials and methods. Reaction products from of phosphodiesterase. In the absence of the enzyme the radio- incubations with or without phosphodiesterase were separated by t.l.c. on activity remained bound to the protein and did not migrate polyethyleneimine cellulose. Spots corresponding to 5'-AMP were extracted and counted as described in Materials and methods. from the origin. 5'-AMP seems to be the sole product released under these conditions as no radioactivity was found any- where else on the t.l.c. plate. It was therefore concluded that cubation with cholera toxin is likely to result from an ADP- the labeling of tubulin by [32p] or [8-3H]NAD following in- ribosylation reaction. 183 Y.Amnr-Zaltsman et al. <-IEF 1 2 3 4 5 _ _

. _ x_.::* m :Vg..-

_t gE. _|. :2:2:' ..... :.:_ W.

s;.: .' .. .: ....: *::.

:r " '.-,

EX .. ...,:,. 4_ .j._.

,ffi,... Jig. 6. Iwo dimensional gel electrophoresis of proteins extracted fromn *: ', j' '' permeated neuroblastoma cells incubated with [32p]NAD and cholera toxin. The reaction was carried out as desribed in the legend to Figure 5 lane 5. ::. ::. *'-F The labeled proteins were subjected to isoelectric focusing. The resulting *.s: ,. isoelectric focusing gels were separated (second dimension) by SDS-PAGE (8 - 12% polyacrylamide gradient, 10 cm long). The arrow indicates the 2.. .i: position of tubulin migration. .s.. aQW ,.

..^,,-,.: _..... 1) revealed many more proteins as compared with the toxin- -.,:. *;.'.''''.'.'Sr,- mediated ADP-ribosylated products (lane 5). The identity of the ADP-ribosylated tubulin was further established by two- dimensional gel electrophoresis (Figure 6) and limited pro- t., tease digestion (not shown).

.i;:..*32 ' Discussion :: r.: .:M .g It was demonstrated in this study that incubation of sE., ' ':::'t':y': purified rat braln tubulin with radiolabeled NAD and cholera toxin leads to the labeling of a and (3 tubulin subunits (Figures 1, 4). High mol. wt. microtubule-associated proteins MAPI and MAP2 are also labeled by this procedure (Figure fig. 5. SDS-PAGE of proteins extracted from permeated neuroblastoma 1). The finding that radiolabeled5'(-AMP may be released cells incubated with [32P]NAD and cholera toxin. The reaction was carried out as described in Materials and methods. Lane 1, crude extract of neuro- from such modified tubulin preparations by snake venom blastoma cells metabolically labeled with [35S]methionine; lane 2, permeated phosphodiesterase (Table I) supports the view that protein cells incubated with [32P]NAD in the absence of cholera toxin; lane 3, mol. labeling results from ADP-ribosylation catalyzed by cholera wt. markers stained with Coomassie brilliant blue (130 K, 92.5 K, 69 K, 46 toxin. This reaction seems to involve at least four tubulin K, 30 K); lane 5, permeated cells incubated with [32P]NAD and cholera forms as toxin; lane 4, Coomassie brilliant blue staining of the proteins present in (Figure 2), judged by autoradiography. It should be lane 5. Autoradiograms are presented in lanes 1, 2, and 5. T represents the bor in mind that each of these spots (Figure 2), represents a position of migration of tubulin from neuroblastoma B104 cells. mixture of several isotubulins (Gozesabl,et 1979) which are not well-resolved by this procedure. The labeled bands seem To determine whether tubulin might serve as an en- to represent modified proteins and not poly(ADP-ribose) dogenous substrate for ADP-ribosylation catalyzed by since treatment with S. aurwu protease resulted in their cholera toxin in intact cells, permeated rat neuroblastoma degradation. By electrophoretic mobility on one- and two- B104 cells were used. Cells were treated with lysolecithin dimensional gel electrophoresis and by peptide mapping we (Miller et al., 1978) to permit penetration of labeled NAD identified the labeled proteins as a and (3 tubulin. molecules. More than 90%7o of the cells stained with trypan It was not surprising to find that the extra negative charges blue after exposure to lysolecithin at 4°C but retained their added to the polypeptide chalns by the ADP-ribose moeity gross cellular morphology. Following incubation for 45 min resulted in an acid shift in mobility of the modified tubulin with [32P]NAD and cholera toxin at 30°C, the permeated cells subunits (Figure 2). The rather uniform acid shift in pI (A pl were collected, washed once with buffer, and the proteins = 0.18) of the four tubulin forms indicates that a similar subjected to SDS-PAGE. Figure 5 shows that one of the ma- degree of modification of these proteins has probably taken jor proteins labeled in this experiment comigrated with neuro- place. The a and tubuein subunits, as well as MAPI and blastoma tubulin (lane 5). In the absence of the toxin in the MAP2, may thus be considered as potential targets for reaction mixture no labeling of the cell tubulin was observed cholera toxin action in vivo. (lane 2). It should be noted that the Coomassie brilliant blue It was shown that several proteins can serve as substrates staining of the neuroblastoma protein (lane 4) and a crude ex- for the ADP-ribosylation reaction (Watkins et al., 1980). We, tract of cells metabolically labeled with [35S]methionine (lane therefore, chose a permeated neuroblastoma cell system to 184 ADP-Ribosylation of mkcrotubule protein

of this reaction. Under these condi- Johnson et al., 1978, as previously described (Amir-Zaltzman et al., 1980). examine the specificity The labeling reaction in a fial volume of 100 Il contained 50 mM potassium tions only a few of the cell proteins were labeled and one of phosphate pH 7.2, 0.5 mM ATP, 0.1 mM GTP, 20 mM thymidine, 20mM the major substrates of this reaction was tubulin (Figure 5). A arginine, 5 mM ADP-ribose, 10 MM [3P]NAD or [8-3H]NAD, 100-200 itg small proportion of the tubulin subunits (Zisapel and Lit- acceptor proteins, and 22 Ag cholera toxin. Incubation was for 20 min at tauer, 1979; Gozes and Littauer, 1979) have been reported to 30°C. The reaction was terminated by addition of ice cold 8007 acetone and be associated with cell membranes. It is, therefore, not the material was stored at - 200C. unlikely that these proteins may have been labeled in the Electrophoresis and peptide mapping the acetone-precipitated proteins were centrifuged course of cholera toxin dependent ADP-ribosylation of cell Prior to electrophoresis, at 20 OOOg for 15 min and the pellets were taken up in 3% SDS containing 5% membranes in vitro. Indeed, studies performed with cholera mercaptoethanol and incubated for 1 h at 60°C. SDS-PAGE was performed toxin in relation to the GTP-binding protein associated with on slab gels (1.6 mm x 13 cm x 18 cm) according to Laemmli, 1970 with adenylate cyclase have reported the presence of membrane- 7.5% -15%or 15% - 20% polyacrylamide gradients, or with constant 8% bound 52-53 K labeled protein bands (Johnson, 1978; Hud- polyacrylamide. Separation was accomplished at constant voltage (100-120 of the present findings V) for 14-16 h. Gels were stained with Coomassie brilliant blue, dried, and son and Johnson, 1980). In the light exposed to X-ray films (Agfa Gevaert RP.2). Gels containing 3H-labeled pro- the relationship of these proteins to microtubules may be teins were treated with sodium salicylate prior to autoradiography tested. Interestingly cholera toxin catalyzed ADP-ribosylation (Chamberlin, 1979). of the membrane-bound GTP-binding protein of pigeon Protease digestion of excised tubulin bands from 8%o gels was performed and according to Cleveland et al., 1977, with slight modifications as previously erythrocytes (Enomoto and Gill, 1980) or of purified 45 using 15% et described (Gozes and Littauer, 1978; Gozes et al., 1979) -20%7o 52 K subunits of this protein from rabbit liver (Northrop polyacrylamide gradient and 25 ng S. aureus V8 protease/slot. Isoelectric- al., 1980) requires the presence of yet another macro- focusing-electrophoresis of protein samples was performed as previously molecular factor. In contrast, the exogenous addition of such described (Gozes and Littauer, 1978). a factor for ADP-ribosylation of purified rat brain tubulin by Phosphodiesterase treatment of ADP-ribosylated proteins cholera toxin seems not to be essential. The same is also true Identification of radioactive products released from "ADP-ribosylated" for purified rat brain actin consisting of and -y subunits tubulin was carried out by digestion with snake venom phosphodiesterase ac- with the following modifications: labeled which we found in this study to be readily labeled by NAD in cording to Watkins et al., 1980, tubulin bands were excised from non-fixed non-stained 8% gels, cut into smnall the presence of cholera toxin (data not shown). pieces, and soaked for 12 h in 1.5 ml 10 mM Tris-acetate pH 7.0 at room It is interesting to note that endogenous ADP-ribosyltrans- temperature with light shaking. Bovine serum albumin (100 Ag) was added to ferase activity is present in animal cells (Moss and Vaughan, each tube before precipitation with ice cold 5% trichloroacetic acid. The pellet 1978). It is therefore possible that tubulin and MAPs may be was extracted twice with dry ethyl ether. Washed precipitates were suspended subjected to snake venom phosphodiesterase as describ- by such enzymes in vivo. The physiological in reaction buffer and ADP-ribosylated ed by Watkinset al., 1980. Samples of the reaction mixture were loaded on to implications of ADP-ribosylation of microtubule proteins thin layer polyethyleneimine-cellulose plates and chromatographed in the may be viewed as potential post-translational modifications presence of 0.01 zmol of unlabeled carrier NAD and 5'-AMP with 0.25 M that regulate the structural organization of microtubule pro- LiCI at room temperature. Spots were identified by u.v. light. Extraction and teins and their assembly (Littauer et al., 1980). These also in- counting of radioactive nucleotides were performed as described earlier clude the control of interaction of microtubule proteins with (Salomon and Rodbell, 1975). other cytoskeletal networks, membranes, and elements of ADP-ribosylation ofpermeated neuroblastoma cells Rat neuroblastoma B104 cell line was kindly provided by D.Schubert of the chromatin. Salk Institute, La Jolla, CA. Stock cultures were maintained in Dulbecco- Vogt modification of Eagle's medium (DMEM) supplemented with 8% fetal Materials and methods calf serum in a humidified atmosphere of 5%o C02/95%o air at 37°C and sub- cultured every 4-6 days with 0.25% pancreatin in DMEM. Cells were grown [a32P]ATP (10-40 Ci/mmol), [8-3H]ATP (10-40 Ci/mmol) and a mix- saline The Radio- to the confluent state, washed three times with phosphate buffered ture of 14C-methylated protein markers were purchased from temperature, incubated in a solution containing 0.25 mM toxin was obtained from (PBS) at room chemical Center, Amersham. Cholera 0.125 mM EGTA, 30 mM Tris and 0.25 M sucrose, pH 7.8 for 10 min was from Boehringer. (3-nicotin- EDTA, Schwarz/Mann. NAD pyrophosphorylase at 37°C. The detached cells were collected by centrifugation for 5 min at 500 amide mononucleotide (NMN) and lysolecithin were from Sigma. S. aureus suspended in cold 150 mM sucrose, 80 mM Snake venom g, washed once with PBS and V8 protease (36-900) was obtained from Miles laboratories. 5 mM potassium phosphate (7.4) (solution B) at Phosphocellose P1 was KCI, 35 mM Hepes (7.4), phosphodiesterase was obtained from Worthington. The were chilled on ice for 5 min, I/IO volume of were of analytical grade. 1-3 x 107 cells/ml. cells purchased from Whatman. All other reagents 1 mg/ml lysolecithin in solution B was added and the suspension incubated Isolation of microtubutes and purification of tubulin for 1 min at 4°C (Miller, 1978). The suspension was then warmed to 37°C. Microtubules were isolated from 30-day-old male rat brains by two cycles Aliquots of 0.1 ml of the permeated cells were incubated for 45 min at 30°C in of assembly-disassembly according to Shelanski et al., 1973. Microtubule a reaction mixture (final volume 0.3 ml) containing: 10 M [32P]NAD (1-5 x pellets were stored at -70°C. Tubulin was purified from assembled micro- 107 c.p.m.), 50 mM potassium phosphate buffer, and, when indicated, 48 Ag tubules by chromatography on phosphocellulose with a slight modification of activated cholera toxin. The suspension was then centrifuged for 5 min at 500 the method of Weingarten et al., 1975. Microtubules were depolymerized in g in the cold, washed once with 0.1 ml of solution B, and the pellet dissolved ice cold 0.025 M morpholino-ethansulfonic acid pH 6.6, containing 1 mM in a solution containing 9.5 M urea, 2% NP40, and 5%o (3-mercaptoethanol. EDTA, 0.5 mM MgCl2, and 1 mM dithiothreitol (Buffer I) for 30 min at 4°C. The solution was then centrifuged at 100 000 g for 30 min in the cold. The Acknowledgements fraction (12.5 mg protein) was applied onto a 5-ml bed phos- supernatant for helpful discussions and Mrs. M.Kopelowitz column previously equilibrated with Buffer I. Tubulin was eluted We thank Dr. H.R.Lindner phocellulose work was supported by grants to 10 ml of Buffer I at a flow rate of 10 ml/h. The column was washed with for excellent secretarial assistance. This with Foundation and Population Council, Inc., NY and the 50 ml of Buffer I and the MAP fraction was then eluted with 10 ml of 0.8 M H.R.L. by the Ford Foundation, to Y.S. by the U.S.-Israel Binational Science Foun- NaCl in Buffer I. The two protein fractions were either used immediately or Rockefeller dation (BSF), Jerusalem, and to U.Z.L. by the Muscular Distrophy Associa- frozen at - 70°C. Protein determination was according to Bradford, kept of the Charles W. and Tillie Lubin Career 1976, using bovine serum albumin as standard. tion. Y.S. is the incumbent Development Chair. ADP-ribosylation reaction Preparation of [32P]NAD from [a32PIATP and [8-3H]NAD from [8- References 3H]ATP was carried out according to Cassel and Pfueffer, 1978, or was purchased from The Radiochemical Center, Amersham. Activation of cholera Amir-Zaltsman,Y., Ezra,E., Walker,N., Lindner,H.R., and Salomon,Y. toxin with 20 mM dithiothreitol was according to Moss et at., 1976. Labeling (1980) FEBS Lett., 122, 166-170. of proteins with [32P]NAD or [8-3H]NAD was performed according to Bradford,M.M. (1976) Anal. Biocnem., 72, 248-254. 185 Y.Amir-Ziatsman et al.

Cassel,D., and Pfueffer,T. (1978) Proc. Nati. Acad. Sci. USA, 75, Neurotransmitters and Their Receptors, John Wiley and Sons, London, 2669-2673. pp. 547-557. Chamberlain,J.P. (1979) Anal. Biochem., 98, 132-135. Mekalanos,J.J., Collier,R.J., and Romig,W.R. (1979) J. Biol. Chem., 254, Cleveland,D.W., Fischer,S.G., Kirschner,H.W., and Laemmli,U.K. (1977) 5849-5854. J. Biol. Chem., 252, 1102-1106. Miller,M.R., Castellot,J.J.,Jr., and Pardee,A.B. (1978) Biochemistry Enomoto,K., and Gill,D.M. (1980) Proc. Nati. Acad. Sci. USA, 75, (Wash.), 17, 1073-1080. 3621-3624. Moss,J., Manganiello,V.C., and Vaughan,M. (1976) Proc. Natl. Acad. Sci. Gill,D.M., and Meren,R. (1978) Proc. Natl. Acad. Sci. USA, 75, USA, 73, 4424-4427. 3050-3054. Moss,J., and Vaughan,M. (1978) Proc. Natl. Acad. Sci. USA, 75, Gozes,I., and littauer,U.Z. (1978) Nature, 276, 411-413. 3621-3624. Gozes,I., and Littauer,U.Z. (1979) FEBS Lett., 99, 86-90. Northrop,J.K., Sternweis,P.C., Smigel,M.D., Schleifer,L.S., Ross,E.M., Gozes,I., Saya,D., and Littauer,U.Z. (1979) Brain Res., 171, 171-175. and Gilman,A.G. (1980) Proc. Nati. Acad. Sci. USA, 77, 6516-6520. Hayaishi,O., and Veda,K. (1977) Annu. Rev. Biochem., 46, 95-116. Pappenheimer,A.M.Jr. (1977) Annu. Rev. Biochem., 46, 69-94. Hinz,H., Adamietz,P., Bredehorst,R., and Wielckens,K. (1978) in Weber,G. Salomon,Y., and Rodbell,M. (1975) J. Biol. Chem., 250, 4253-4260. (ed.) Advances in Enzyme Regulation, Vol. 17, Pergamon Press, Oxford/ Shelanski,M.L., Gaskin,F., and Cantor.C.R. (1973) Proc. Natl. Acad. NY, pp. 195-212. Sci. USA, 70, 765-768. Hudson,T.H., and Johnson,G.L. (1980) J. Biol. Chem., 255, 7480-7486. Watkins,P.A., Moss,J., and Vaughan,M. (1980) J. Biol. Chem., 255, Johnson,G.L., Kaslow,H.R., and Bourne,H.R. (1978) J. Biol. Chem., 3959-3963. 253, 7120-7123. Weingarten,M.D., Lockwood,A.H., Huo,S.Y., and Kirschner,M.W. (1975) Laemmli,U.K. (1970) Nature, 227, 680-685. Proc. Natl. Acad. Sci. USA, 73, 1858-1862. Littauer,U.Z., de Baetselier,A., Ginzburg,l., and Gozes,I. (1980) in Zisapel,N., and Littauer,U.Z. (1979) Eur. J. Biochem., 95, 51-59. Littauer,U.Z., Dudai,Y., Silman,I., Teichberg,V.I., and Vogel,Z. (eds.),

186