Purification of Bovine Brian Tubulins

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1990 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 265, No. 3, Issue of January 25, pp. 1794-1’799,199O Printed in U.S.A.

Increased Microtubule Assembly in Bovine Brain Tubulin Lacking the Type III Isotype of ,&Tubulin*

(Received for publication, June 26, 1989)

Asok Banerjee, Mary Carmen Roach, Phyllis Trcka, and Richard F. Ludueiia$ From the Department of Biochemistry, the University of Texas Health Science Center, San Antonio, Texas 78284

Tuhulin, the major constituent protein of microtu- bules, is a heterodimer of cr and B subunits. Both (Y and B exist in multiple isotypic forms. It is not clear if different isotypes perform different functions. In order to approach this question, we have made a monoclonal antibody specific for the &u isotype of tubulin. This particular isotype is neuron-specific and appears to be phosphorylated near the C terminus. We have used immunoaffinity depletion chromatography to prepare tubulin lacking the BIII subunit. We find that removal of the ,&u isotype results in a tubulin mixture able to assemble much more rapidly than is unfractionated tubulin when reconstituted with either of the two mi- crotubule-associated proteins (MAPS), tau or MAP 2. Our results suggest that the different isotypes of tu- bulin differ from each other in their ability to poly- merize into microtubules. We have also found that the anti-& antibody can stimulate microtubule assembly when reconstituted with tubulin and either tau or MAP 2. When reconstituted with tubulin lacking the & isotype, the antibody causes the tubulin to polymerize into a polymer that is a microtubule in the presence of MAP 2 and a ribbon in the presence of tau.

Microtubules are cylindrical organelles which perform a great variety of functions in eukaryotic cells; among the processes in which they participate are mitosis, ciliary and flagellar motility, intra-cellular transport, and the generation and maintenance of cell shape (Dustin, 1984). Microtubules are composed of a 100 kDa protein, tub&z, which consists of two 50-kDa subunits designated o( and p (Dustin, 1984). In addition to tubulin, microtubules are attached to a variety of non-tubulin proteins designated microtubule-associated pro- teins (MAPS)’ (Vallee and Bloom, 1984). Because of the many functions which microtubules perform, there has long been speculation that different microtubule-mediated functions may require different forms of tubulin (Behnke and Forer, 1967; Fulton and Simpson, 1976). Recent studies have dem- onstrated that both (Y- and @-tubulin exist as multiple isotypes encoded, in mammals or birds, by a set of six 01 and seven /j’

* This work was supported by Grants GM23476 from the National Institutes of Health and AQ-0726 from the Robert A. Welch Foun- dation (to R.F.L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed. 1 The abbreviations used are: MAPS, microtubule-associated pro-

teins; &, type I isotype of /3-tubulin; &, type II isotype of P-tubulin; &,I, type III isotype of fl-tubulin, &v, type IV isotype of fl-tubulin; @v, type V isotope of fl-tubulin (this is the nomenclature of Sullivan and Cleveland, 1986); MES, 4-morpholineethanesulfonic acid; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid.

genes (Villasante et al., 1986; Monteiro and Cleveland, 1987). There does not appear to be any tissue in which all of these isotypes are expressed, but in mammalian brain, for example, there appear to be five (Y and five @ isotypes (Villasante et al., 1986; Sullivan and Cleveland, 1986). This in turn suggests that there may exist in the brain as few as five or as many as 25 different c@ heterodimers. It is not yet clear, however, that the different tubulin heterodimers are functionally different. On the one hand, various experiments, all incidentally involv- ing non-neuronal cells, have found that functionally different microtubules can contain multiple isotypes of either cy- or p- tubulin; even when various manipulations are used to intro- duce heterologous tubulin isotypes into the cells, there does not appear to be a change in any of the functional properties of the cell’s microtubules (Bond et al., 1986; Lewis et al., 1987; Lopata and Cleveland, 1987; Joshi et al., 1987; Gu et al. 1988; Lewis and Cowan, 1988). These observations would imply that in these cells, the differences among the isotypes do not have a functional significance. On the other hand, Luduena et al. (1985) have found that assembly of brain, but not of erythrocyte, tubulin is inhibitable by a low concentration of a non-physiological sulfhydryl alkylating agent, indicating that different isotypes of tubulin can behave differently in vitro. Along the same lines, there is evidence that different LY- tubulin isotypes occur in different regions within the same neuron, suggesting that they may perform different functions (Hajos and Gallatz, 1985). Also, Joshi and Cleveland (1988) have shown that, during neurite outgrowth in neuroblastoma cells, the @I- and &-tubulin isotypes are incorporated into microtubules in preference to the & isotypes. Finally, Gard and Kirschner (1985) and Luduena et al. (1988) have found that only one /3-tubulin isotype, designated&l, becomes phos- phorylated at a serine residue near the C terminus upon differentiation of neuroblastoma cells; the other p isotypes do not appear to become phosphorylated under these conditions. It is probably significant that the C-terminal tryptic peptide of PIIr contains two serines that are not present in the other neural p isotypes.

In order to test whether the isotypes are functionally dis- tinct, it is first necessary to demonstrate that they have different functional properties in uitro. For that it is necessary to separate them in a functional state. To attain this end, we are using isotype-specific monoclonal antibodies. We have chosen to work with brain tubulin because the brain is the tissue richest in tub&n isotypes and because such evidence as there is which suggests that different tubulin isotypes are functionally different involves brain tubulin. We are concen- trating on p rather than cy because the sequences of the p isotypes are more variable than those of the (Y isotypes. We have previously reported the preparation of a monoclonal antibody specific for the brain /3-tubulin isotype &, which constitutes 58% of cerebral P-tubulin (Banerjee et al., 1988). We here describe the preparation of a monoclonal antibody

1794

by guest on A ugust 8, 2017

http://w w

w .jbc.org/

D ow

nloaded from

Removal of aPIII-Tub&n Promotes Assembly

specific for the ,&l isotype, which constitutes 25% of cerebral @-tubulin and which appears to be entirely restricted to the brain (Little, 1979; Luduena et al., 1982; Sullivan and Cleve- land, 1986). We have used this monoclonal antibody to pre- pare brain tubulin free of the Plrr isotype. We have found that tubulin lacking ,&n assembles in uitro into microtubules twice as rapidly as does normal brain tubulin. Our results suggest that the isotypic composition of tubulin strongly influence its assembly properties.

volume of 50% (w/v) sucrose in assembly buffer. 10 ~1 of this diluted mixture was placed on a collodion- and carbon-coated copper grid (200-mesh) and was allowed to adsorb for 1 min. The grid was successively washed with four drops each of the following: (a) 1 mg/ ml cytochrome c in water, (b) water, and (c) 1% uranyl acetate in water. Excess stain was soaked off with a blotting paper. The grids were air-dried and examined under a Jeol 1OOCX electron microscope at an accelerating voltage of 60 kV. Protein determination was done according to Lowry et al. (1951).

EXPERIMENTAL PROCEDURES

Materials-The peptide Cys-Glu-Ser-Glu-Ser-Glu-Gln-Gly-Pro- Lys-OH was synthesized by the BioSearch Corporation, San Rafael, CA. Except for the amino-terminal cysteine, the peptide has a se- quence identical to that of the last amino acids in human &tubulin (Sullivan and Cleveland, 1986). Grids for electron microscopy were from Ted Pella. Inc. Tustin. CA. All other materials were uurchased as described previously (Luduena et al., 1982; Banerjee et il., 1988).

Partial Purification of Tubulin Isotypes-Tubulin was purified from bovine cerebra by chromatography on phosphocellulose as described by Fellous et al. (1977). The purified tubulin was subjected to chro- matography on an immunoaffinity column containing the antibody JDR.3B8, which binds specifically to the pii and pi isotypes of tubulin (Banerjee et al., 1988). The unbound tubulin was collected and the bound tubulin was then obtained by eluting the column with 3.5 M KI as previously described (Banerjee et al., 1988). Based on previous results (Banerjee et al., 1988), the unbound fraction should contain (Y, prri and /3rv and the bound fraction should contain 01, pr and &. Both fractions were reduced and carboxymethylated and subjected to preparative electrophoresis (Laemmli, 1970) on gels containing 5.5% polyacrylamide. The gels were soaked in 4 M sodium acetate in order to visualize the bands, the bands were cut out and the protein extracted therefrom (Higgins and Dahmus, 1979). From the gel of the unbound fraction, (Y, @iii, and piv were obtained by taking advantage of the fact that & has a unique electrophoretic mobility. From the gel of the bound fraction, o( was purified as well as a mixture of fir and PII.

Preparation of Monoclonal Antibody against Type III P-Tubulin- The peptide CESESQGPK was conjugated with bovine serum albu- min and m-maleimidobenzoyl-N-hydroxysuccinimide and injected into BALB/c mice; hybridoma cell lines were prepared as previously described (Banerjee et al., 1988).

The selection procedure was as follows. Clones were initially screened against phosphocellulose-purified tubulin; 22 positive wells (out of 384) were found. These 22 positives were further screened against purified 01, &r, and &, as well as the mixture of fir and ,&I. From the 22 positives, four were selective for subcloning on irradiated thymocytes (McKearn, 1980). 80 wells tested positive against tubulin from the unbound fraction of the immunoaffinity column (this tu- bulin would have consisted of (Y, /j’ 111, and prv) at a concentration of 2 pg/ml. Of these, 48 clones were tested against prv and cu; nine single- cell clones tested negative against piv and a and registered a very low positive to the pr/flrr mixture. Cell-free supernatants of six cell lines were tested by immunoprecipitation using protein A-agarose. These six also tested negative against peptides corresponding to the C termini of the @rv and flv subunits and tested positive to the peptide corresponding to the C terminus of flrrr as well as to a mixture of&r- and @Iv-tubulins. The cell line SDL.SDlO bound best to protein A and was selected for growing in roller bottles and for subsequent purification by chromatography on protein A-agarose as previously described (Baneriee et al., 1988). The antibody SDL.3DlO tested positive against isolated &tubulin; neither a nor @rr showed reactiv- ity. The isotype of SDL.3DlO was determined to be IgG2b.

Electrophoresis and Electroblotting-Polyacrylamide gel electro- phoresis was done on 5.5% gels containing 0.1% sodium dodecyl sulfate (Laemmli. 1970). Electroblottine was carried out at 30 V overnight at 4 “C’and Western analysis-performed as described by Banerjee et al. (1988).

Other Techniques-Microtubule assembly was monitored at 37 “C in a Gilford model 250 spectrophotometer equipped with an automatic cuvette programmer. Cuvette temperature was controlled by using a Gilford thermoset. All assembly experiments were done in the follow- ing buffer: 0.1 M MES, pH 6.4, 1 mM EGTA, 0.1 mM EDTA, 1 mM GTP, and 0.5 mM MgCl,. Electron microscopy was done according to Olmsted and Borisy (1973) with little modification as described below. After assembly, the assembled mixture was diluted with an equal

RESULTS

Chromatography of Phosphocellulose-purified Tubulin on the Antibody-Sepharose Affinity Matrix-An affinity column was prepared by coupling the antibody SDL.3DlO with CNBr- activated Sepharose. Phosphocellulose was passed through the column, and the voided unbound fraction was collected. After washing the column with the running buffer, the bound fraction was eluted with 3 M KI. Both the bound and the unbound tubulin fractions were reduced, carboxymethylated, and subjected to polyacrylamide gel electrophoresis. The re- sults (Fig. la) show that the unbound fraction (U) lacks the p2 band as expected; however, the bound fraction (B) not only contains the p2 band but also a significant & band. The ratio of the fi2 to the 6, bands in the bound fraction was about 60~40.

Portions of the same gel shown in Fig. la were reserved for immunoblotting against the anti-& and the anti-&i antibod- ies. The results with the anti-& antibody (SDL.3DlO) (Fig. lb) showed the presence of @iii-tubulin in the bound fraction as well as in unchromatographed tubulin (PC), but not in the

a b C

PC u B PC U B PC U B

FIG. 1. Fractionation of bovine brain tubulin on an affinity column containing covalently linked SDL.3D10, the mono- clonal antibods saecific for the &,r isotype. Tubulin that had been purified by-chromatography on phosphocellulose was chromat- ographed on a Sepharose column containing the antibody SDL.3DlO. Bound material was eluted with 3 M KI. All samples were reduced and carboxymethylated and subjected to electrophoresis on a 5.5% polyacrylamide gel (Laemmli, 1970). Three identical gels were run. Five wg of each sample was loaded onto the gels. One set (a) was reserved for staining with Coomassie Blue. Sets b and c were blotted onto nitrocellulose membranes which were treated first with either SDL.BDlO, the monoclonal antibody specific for the flrrr isotype (b), or JDR.3B8, the monoclonal antibody specific for the flii isotype (c). The blots were then treated with ‘*“I-labeled sheep anti-mouse anti- body, prior to being dried and visualized by autoradiography. PC, phosphocellulose-purified tubulin that was applied to the column; U, the fraction of tubulin that did not bind to the column; B, the fraction of tubulin bound to the column and was eluted with KI. Note: the designations pi and & were given to the two fl-tubulin bands by Little (1979). The pin isotype migrates in the /& band, and the other /3 isotypes migrate in the fll band. The p isotypes migrating in the individual bands are designated on the right side of the figure.

by guest on A ugust 8, 2017

http://w w

w .jbc.org/

D ow

nloaded from

1796 Removal of a~III-Tubulin Promotes Assembly

unbound fraction, indicating that the column had removed virtually all of the Pin-tubulin from the chromatographed tubulin. Analysis with the anti-&i antibody (JDR.3B8) showed that both the bound and the unbound fractions con- tained Pii-tubulin (Fig. 1 c) .

It is not clear why the anti-& antibody is less specific as part of an immunoaffinity column than on an ELISA assay. One possibility is that the antibody has a weak affinity for Pii-tubulin and perhaps for @iv-tubulin as well. Although the C-terminal sequence of pIi1 differs significantly from those of the other fl-tubulin isotypes, all these sequences are rich in glutamate, a similarity which could conceivably allow for some weak interactions between the anti-& antibody and the other isotypes (Sullivan and Cleveland, 1986). On the ELISA where the concentrations of the tubulin and the antibody are both low, only the specific binding to &ii is seen; in contrast, in the affinity column, where both the antibody and the tubulin are likely to be present in very high concentrations, it is possible that some nonspecific binding could occur. Another possibility is that the antibody, when it binds to @iii-tubulin, alters its properties in such a way so as to cause an interaction between the &in dimer and other dimers such as a& By this model, the &i-tubulin would bind indirectly to the affinity column. In order to see whether the &i antibody column binds other isotypes directly, we passed /3in-depleted tubulin through that column. It was found that the column can still bind to tubulin, indicating that the piii antibody column binds directly but nonspecifically to isotypes other than pm. The bound tubulin was eluted and subjected to electrophoresis on polyacrylamide gels followed by immunoblotting using monoclonal antibodies to either &i or &v; both of these isotypes were present. In an analogous experiment, samples of either unfractionated tu- bulin and PIrr-depleted tubulin, at concentrations of 2.0 mg/ ml, were incubated with the antibody (0.22 mg/ml) for 30 min at 0 “C and then treated with protein A-agarose (16 mg/ml) for another 30 min, prior to centrifugation. The resulting immunoprecipitate was rinsed three times and then separated from the protein A-agarose, reduced and carboxymethylated, and analyzed on polyacrylamide gels. It was clear that the anti-/3iii antibody could precipitate the pi band from either the unfractionated or &depleted tubulin. Since the p1 band contains the isotypes pi, pii, and @iv, it is apparent that the antibody, even though it binds preferably to @iii can be made to interact with other isotypes as well.

Microtubule Assembly in the Presence of Anti-Pl~~---In order to examine the effect on microtubule assembly of the antibody against Pin-tubulin, the experiment shown in Fig. 2 was per- formed, in which a series of antibody concentrations was added to samples of unfractionated tubulin in the presence of either tau or MAP 2. As can be seen, the antibody enhanced microtubule assembly; the enhancement was observed in the presence of either tau or MAP 2. Increasing antibody concen- tration caused increased polymerization. It is unlikely, how- ever, that the effect could be due simply to the antibody cross- linking the tubulin molecules, since the extent of polymeri- zation was apparently doubled at antibody/tubulin molar ratios of 0.063 (Fig. 2).* This fact also decreases the likelihood that the increased turbidity seen in the presence of antibody is due only to the formation of antibody-tubulin complexes. The structures assembled in the presence of MAP 2 over the full range of antibody/tubulin ratios examined appeared to be normal microtubules (Fig. 3~). Microtubules were formed in the presence of tau as well (Fig. 36), but at high antibody/ tubulin ratios (l:lO), the product assembled in the presence

* Assuming molecular weights for IgG and tubulin of 150,000 and 100,000, respectively.

04 I

Ttme (Man) Time (Man)

FIG. 2. Effect of anti-flu, on the assembly of tubulin. Aliquots (250 ~1) of nhosnhocellulose-Durified tubulin (2.0 mg/ml) were olaced in cuvettes-together with either MAP 2 (0.30 mg/mij (a) or ta; (0.15 mg/ml) (b). Samples in a contained the following concentrations of the monoclonal antibody SDL.3DlO as indicated here: 1, 0 mg/ml; 2, 0.094 mg/ml; 3,0.188 mg/ml; 4,0.282 mg/ml). Samples in b contained the following concentrations of the monoclonal antibody SDL.3DlO as indicated here: 1,O mg/ml; 2,0.141 mg/ml; 3,0.188 mg/ml; 4,0.282 mg/ml; 5, 0.376 mg/ml. The cuvettes were placed in a Gilford spec- trophotometer at 37 “C and polymerization was monitored by turbi- dimetry at 350 nm.

of tau consisted largely of ribbons (Fig. 3~). The antibody- induced enhancement of assembly is consistent with the pos- sibility that the c&n dimer assembles less well than do the other dimers, and the antibody enhances assembly by seques- tering the @iii dimers and preventing them from participating in assembly.

Assembly of Tubulin Depleted of Type ZZZ p-Tubulin-Inter- pretation of the experiment shown in Fig. 2 was complicated by the fact that the antibody was present. In order to see if the polymerization-enhancing effect could be due to the se- questering of the &ii dimers by the antibody, we examined the assembly into microtubules of tubulin depleted of &ii- tubulin by passage through the immunoaffinity column. In these experiments, the &-depleted tubulin was reconstituted with either tau or MAP 2 and compared with a control of unresolved tubulin similarly reconstituted. The results (Fig. 4) show that both the rate and extent of assembly are signif- icantly enhanced in the case of the &ii-depleted tubulin as compared with that of unresolved tubulin. Similar results were obtained with both MAP 2 and tau. Electron microscopic studies showed that the products of assembly from Bin-de- pleted tubulin appeared to be normal microtubules (Fig. 5). In order to see if enhancement of assembly by immunodeple- tion could be an artifact arising from nonspecific binding to Sepharose of a hypothetical inhibitory substance of unknown nature, which would then be removed from the tubulin prep- aration and permit enhanced assembly, the following control experiment was done, in which tubulin was passed through an immunoaffinity column containing bovine serum albumin instead of tubulin. This tubulin did not show any increase in its ability to assemble into microtubules as compared with unfractionated tubulin that had not been passed through the column3

Effect of Anti-&If on the Assembly of Tubulin Depleted of PI,,-Tubulin-One might expect that when the antibody against PrII is added to tubulin depleted of Pin, then there would be no effect on assembly. This is not the case, however. In fact, when the antibody is added to tubulin depleted of &ii,

3 D. Sachdev, A. Banerjee, and R. F. Luduena, unpublished results.

by guest on A ugust 8, 2017

http://w w

w .jbc.org/

D ow

nloaded from

Removal of cuPIf,-Tubulin Promotes Assembly

b

:> Time (Min) Trne (Mtnl

FIG. 4. Polymerization of &,-depleted tubulin. Aliquots (2.50 ~1) of tubulin (2.0 mg/ml) were incubated at :X7 “C in the presence of eit,her tau (0.22 mg/ml) C/r// pnnel) or MAP 2 (0.26 mg/ml) (right panel). Polymerization was measured by turbidimetry as in Fig. 1. The ability to polymerize was compared between phosphocellulose- purified tubulin that had not been further fractionated (0) and tubulin that had been depleted of the /jr,, isotype by immunoaffinity chromatography (A).

1797

a

C

b

FIG. :3. Microtubules assembled from unfractionated tubu- lin in the presence of anti-&u. Phosphocellulose-purified tubulin (2.0 mg/ml) containing the monoclonal antibody SDL3DlO (0.16 ma/ml) was incubated at 37 “C in the presence of either MAP 2 (0.26 mg/ml) to) or tau (0.22 mg/ml) (b). Notice the presence of microtu- bules in these samples. One sample was incubated with tau (0.22 mg/ ml) and a higher concentration of the antibody (0.22 ma/ml) (c). In the presence of higher antibody concentration numerous ribbons were observed cc). Magnification x 139.000.

the extent and the rate of assembly increase greatly. As can be seen in Fig. 6, after 10 min of incubation at 37 “C in the presence of tau and anti-&, the ,&-depleted tubulin reaches a turbidity that is over two times as great as that attained by an identical sample lacking anti-pin. In a similar experiment using MAP 2 instead of tau, the turbidity is also over two times as great as that of the control. In both cases these are attained by antibody/tubulin molar ratios of 0.04, suggesting that the effect is due to an effect of the antibody on the tubulin molecules and not simply to cross-linking of tubulin by the antibody. Morphologically, the polymer produced in the presence of MAP 2 consists entirely of microtubules (Fig. 7~). However, in the presence of tau, in most expe$ments, the polymer consisted entirely of ribbons about 680 A across, consisting of about 13-14 protofilaments (Fig. 76). In one experiment, the polymer obtained in the presence of tau

FIG. 5. Microtubules assembled from &,-depleted tubulin. Samples of tubulin (2.0 ma/ml) that had been depleted of the /jttt isotype by immunoaffinity chromatography were incubated at 37 “C in the presence of either MAP 2 (0.26 mg/ml) (a) or tau (0.22 mg/ ml) (b). Magnification X 139,000.

contained a mixture of microtubules and curled ribbons (Fig. 7c).

DISCUSSION

It is apparent from Fig. 4 that removal of /jlll from unfrac- tionated tubulin causes a great enhancement of the rate and extent of assembly. There are several possible explanations of this observation. First, it may be that pill-tubulin has an intrinsically lower rate of assembly and is unable to assemble to the same extent as do the other fi isotypes. Hence, the presence of this isotype in a mixture of all of the fi isotypes may result in an overall lowered rate of assembly. Removal of

by guest on A ugust 8, 2017

http://w w

w .jbc.org/

D ow

nloaded from

1798 Removal of cuPI,,-Tubulin Promotes Assembly

Time (Mln) Time (Mln)

FIG. 6. Polymerization of &,-depleted tubulin in the pres- ence of anti-&,,. Aliquots (2.50 ~1) of’ tuhulin (2.0 mg/ml) that had heen depleted of the ~jtrl isotype by immunoaff’inity chromatography were incubated at 37 “C in the presence of either MAP 2 (0.30 mg/ ml) fn) or tau (0.15 mg/ml) (b). Samples in o also contained the f’ollowin,v concentrations of the monoclonal antibody SDL3DlO as indicated here: I, 0 mg/ml; 2. 0.094 mg/ml; 3, 0.141 mg/ml; .I, 0.188 mg/ml). Samples in h contained the following concentrations of the monoclonal antibody SDL.SDIO: I. 0. mg/ml; 3, 0.047 mg/ml; 3, 0.070 mg/ml: I, 0.094 mg/ml; 5, 0.117 mg/ml. Tubulin polymerization was monitored hy turbidimetry at 3.50 nm.

this isotype would then permit higher rates and extents of assembly in the depleted tubulin. It is possible that the reason for the different assembly properties of the @rlr isotype is correlated with the fact that @ill is unique among the p isotypes in being phosphorylated near the C terminus (Luduena et al., 1988). Conceivably, the phosphorylation of this isotype may play a role in determining the rate and extent of microtubule assembly. The possibility that tubulin isotypes have an in- trinsic difference in their ability to polymerize is consistent with the observation of Harada et al. (1986) who showed that two isotypes of @-tubulin from mouse lymphoma cells differed greatly in their ability to form microtubules in oitro.

A second possible explanation for the results shown in Fig. 4 is that the @iii differs from the other isotypes in the nature of its interaction with tau and MAP 2. In other words, its intrinsic ability to assemble may be the same as those of the other isotypes, but its ability to have its assembly stimulated by binding to tau and MAP 2 may be less. Littauer et al. (1986) have found that the region in the tubulin sequence which has the highest affinity for both tau and MAP 2 is the sequence GEFEEEG in positions 434-440 in @-tubulin. This is actually the sequence found only in the pi1 isotype. The /Jr\ isotype has a somewhat similar sequence (GEFEEEA) in this position. The corresponding sequence in fllll is EMYEDDE. It is conceivable that this region in &i may bind to tau or MAP 2 in a way which is less productive for stimulating assembly than is the manner in which the other isotypes bind to MAPS. For example, recent work indicates that tau and MAP 2 share a homologous domain which has three imper- fectly repeated segments (Lewis et al., 1988). These segments are postulated to correspond to the tubulin-binding regions with the implicit corollary that MAPS stimulate assembly by linking several tubulin molecules. It is conceivable that, if the @iii isotype is present in a mixture of assembling tubulin molecules, then when one of the putative tubulin-binding segments of tau or MAP 2 binds to @iII, the interaction of the other tubulin-binding segments with other tubulin molecules is weakened. Thus, the presence of fill1 would result in de- creased microtubule assembly compared to a situation where @rrr-tubulm is absent.

A third possibility is that the different isotypes differ from each other in the rate at which they decay, a process in which

C

FIG. 7. Polymers formed from p,,,-depleted tubulin in the presence of anti-PI,,. Samples of tuhulin (2.0 mg/ml) that had been depleted of the $,,t isotype by immunoaf’finity chromatography were incubated at 37 “C in the presence of the antibody SDLJDlO (0.22 mg/ml) and either MAP 2 (0.26 mg/ml) (a) or tau (0.22 mg/ml) (h and c). Only microtubules were seen in the MAP 2-containing sample (a). In one sample containing tau (b) only flat ribbons were seen (h); in another experiment the sample with tau also contained a mixture of microtubules and curled ribbons (c). Magnification X 139,000.

the ability to assemble is lost in a time-dependent fashion (for a discussion of decay see Luduena (1979)). It is conceiv- able that the @ill isotype may decay faster than other tubulin isotypes, and that the presence of “decayed” @ill in a mixture of the other isotypes may “poison” assembly. In such a case, removal of @ill would result in faster assembly.

A fourth possible explanation is that filli does not differ from the other tubulin isotypes in either its ability to assemble or in its stability against decay and that the only reason why its removal results in faster assembly is that isotypically purer tubulin assembles more readily than more heterogeneous tu- bulin. This effect would be particularly marked if the isotype being removed were /jlll since its sequence differs more from that of the other isotypes than the sequences of the other isotypes differ among themselves. In a preliminary experi- ment, we have found that removal of a/$ and @ii dimers results in little difference in the ability to assemble. In this

by guest on A ugust 8, 2017

http://w w

w .jbc.org/

D ow

nloaded from

Removal of a/3rrl-Tub&in Promotes Assembly 1799

case, the remaining tub&n consists of a mixture of &ii and @iv which differ considerably from each other.

Whatever the explanation of this phenomenon, there is one implication that can be drawn from our results, namely, that the isotypic composition of tubulin strongly influences its assembly properties. Since the work of Burgoyne et al. (1988) suggests that the distribution in vivo of the firri isotype differs from that of the other isotypes, meaning that not all micro- tubules contain this isotype, it is likely that the assembly properties in vivo of isotypically distinct microtubules differ considerably. Similarly, the microtubules being used by many laboratories for study of dynamic properties in vitro are ac- tually chimeric structures composed of multiple isotypes whose dynamic properties may differ significantly from those of microtubules in vivo.

dimer also stabilizes side-to-side interactions then one could predict that normal microtubules could form easily provided that either MAP 2 or a& is present.

One unexpected finding is the strong increase in polymeri- zation when &depleted tubulin is incubated with the mono- clonal antibody against type III& The resulting polymer is a microtubule in the presence of MAP 2 and appears to be largely or entirely ribbons in the presence of tau. Since the antigen for this antibody is lacking in a &i-depleted mixture, one might expect that the addition of antibody against @III- tubulin should have no effect. It is probable that the polym- erization is due to an interaction of the antibody with other tubulin isotypes. We have found that the antibody SDL.3DlO interacts weakly with heterologous tubulin isotypes. In the presence of &iI, most of the antibody will interact with &I and not with other isotypes. However, if pii1 is lacking, or if the antibody is present in high concentrations, then the antibody will interact with the other isotypes. It has been shown that removal of the C terminus of tubulin with subtil- isin results in a tubulin which can assemble readily in the absence of MAPS (Serrano et al., 1984a). Hence, it has been postulated that the C terminus acts as an intrinsic inhibitor of assembly (Maccioni et al., 1985) conceivably because of electrostatic repulsion among the many negative charges in this region. MAPS appear to interact with the C terminal region (Serrano et al., 1984b; Littauer et al., 1986). Thus, by this model, one could postulate that one of the roles of the MAPS is to bind to the C terminus and cover or neutralize the negative charges, hence inducing assembly. It may be that the antibody SDL.3DlO acts like a MAP in this respect when it interacts with the other tubulin isotypes.

It is interesting that the polymer takes the form of a microtubule in the presence of MAP 2 and of a ribbon in the presence of tau. It should be recalled that when a large amount of antibody is added to unfractionated tubulin, tubulin rib- bons also form, but only if tau is present. It may be that in the presence of large concentrations of antibody, some of the antibody is interacting with heterologous isotypes. It appears that during polymerization in the presence of the antibody normal microtubules always form if either MAP 2 or &r is present but that if both of them are absent then it is more difficult for normal microtubules to form. It has been sug- gested that tau is involved in stabilizing end-to-end interac- tions between tubulin molecules and that MAP 2 stabilizes side-to-side interactions during microtubule assembly (Scheele and Borisy, 1979). If one postulates that the ol&

REFERENCES Banejee, A., Roach, M. C., Wall, K. A., Lopata, M. A., Cleveland,

D. W., and Ludueiia, R. F. (1988) J. Biol. &em. 263, 3029-3034 Behnke, O., and Forer, A. (1967) J. Cell Sci. 2, 169-192 Bond, J. F., Fridovich-Keil, Pillus, L., Mulligan, R. C., and Solomon,

F. (1986) Cell 44, 461-468 Burgoyne, R. D., Cambray-Deakin, M. A., Lewis, S. A., Sarkar, S.,

and Cowan, N. J. (1988) EMBO J. 7,2311-2319 Dustin, P. (1984) Microtubules, 2nd Ed., Springer-Verlag, Berlin Fellous, A., Francon, J., Lennon, A.-M., and Nunez, J. (1977) Eur. J.

Biochem. ‘78, 167-174 Fulton, C., and Simpson, P. A. (1976) in Cell Motility (Goldman, R.,

Pollard, T., and Rosenbaum, J., eds), pp. 987-1005, Cold Spring Harbor Publications, Cold Spring Harbor, NY

Gard, D. L., and Kirschner, M. W. (1985) J. Cell Biol. 100, 764-774 Gu, W., Lewis, S. A., and Cowan, N. J. (1988) J. Cell Biol. 106,2011-

2022 Hajos, F., and Gallatz, K. (1985) Histochemistry 82,491-496 Harada. F.. Matsuzaki. F.. and Yahara. I. (1986) EXD. Cell Res. 166.

265-278 ‘ , ,. ,.

Higgins, R. C., and Dahmus, M. E. (1979) Anal. Biochem. 93, 257- 260

Joshi, H. C., and Cleveland, D. W. (1988) J. Cell Biol. 107,727a Joshi, H. C., Yen, T. J., and Cleveland, D. W. (1987) J. Cell Biol.

105,2179-2190 Laemmli, U. K. (1970) Nature 227,680-685 Lewis, S. A., and Cowan, N. J. (1988) J. Cell Biol. 106,2023-2033 Lewis, S. A., Gu, W., and Cowan, N. J. (1987) Cell 49,539-548 Lewis, S. A., Wang, D., and Cowan, N. J. (1988) Science 242, 936-

939 Littauer. U. Z.. Giveon, D., Thierauf. M.. Ginzbure. I., and Ponstinzl,

H. (1986) P;oc. Nat.‘Ac&. Sci. U.‘S. A. 83, 7162-7166 - Little. M. (1979) FEBS L&t. 108.283-286 Lopaia, M: A., and Cleveland, D. W. (1987) J. Cell Biol. 105, 1707-

1720 Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951)

J. Biol. Chem. 193,265-275 Luduena, R. F. (1979) in Microtubules (Roberts, K., and Hyams, J.,

eds) pp. 65-116, Academic Press, London Luduena, R. F., Roach, M. C., Trcka, P. P., Little, M., Palanivelu,

P., Binkley, P., and Prasad, V. (1982) Biochemistry 21,4787-4794 Luduena, R. F., Shooter, E. M., and Wilson, L. (1985) J. Biol. &em.

252,7006-7014 Luduena, R. F., Zimmermann, H.-P., and Little, M. (1988) FEBS

Z&t. 230,142-146 Maccioni, R. B., Serrano, L., and Avila, J. (1985) BioEssays 2, 165-

169 McKearn, T. J. (1980) in Mono&ma1 Antibodies (Kennett, R. H.,

McKearn. T. J.. and Bechtol, K. B.. eds) D. 374. Plenum Publishina Co., New ‘York ’

._

Monteiro, M. J., and Cleveland, D. W. (1988) J. Mol. Biol. 199, 439- 446

Olmsted, J. B., and Borisy, G. G. (1973) Biochemistry 12,4282-4289 Scheele, R. B., and Borisy, G. G. (1979) in Microtubules (Roberts. K.,

and Hyams, J., eds) pp. 175-254, Academic Press, London Serrano, L., de la Torre, J., Maccioni, R. B., and Avila, J. (1984a)

Proc. Natl. Acad. Sci. U. S. A. 81,5989-5993 Serrano, L., Avila, J., and Maccioni, R. B. (1984b) Biochemistry 23,

4675-4681 Sullivan, K. F., and Cleveland, D. W. (1986) Proc. Natl. Acad.

Sci. U. S. A. 83,4327-4331 Vallee, R. B., and Bloom, G. S. (1984) Modern Cell Biol. 3, 21-75 Villasante, A., Wang, D., Dobner, P., Dolph, P., Lewis, S. A., and

Cowan, N. J. (1986) Mol. Cell. Biol. 6, 2409-2419

by guest on A ugust 8, 2017

http://w w

w .jbc.org/

D ow

nloaded from

A Banerjee, M C Roach, P Trcka and R F Ludueña isotype of beta-tubulin.

Increased microtubule assembly in bovine brain tubulin lacking the type III

1990, 265:1794-1799.J. Biol. Chem.

http://www.jbc.org/content/265/3/1794Access the most updated version of this article at

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/265/3/1794.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on A ugust 8, 2017

http://w w

w .jbc.org/

D ow

nloaded from