Microtubule protein phosphorylation in neuroblastoma cells and neurite growth

During brain development, undifferentiated neuroepithelial cells are converted into mature, differentiated neurones that eventually become interconnected through functional synapses. Among the most remarkable events of neuronal differentiation are process extension and the assumption of a highly characteristic cellular morphology.Clonal cell lines derived from tumour cells of neuronal origin have been widely used as model systems to study the earliest events of neuronal morphogenesis. Neuroblastoma cell cultures have been considered a particularly suitable model in this respect. Neuroblastoma tumours usually arise from neural crest-derived precursors of sympathetic neurones that continue proliferating and show a defective differentiation in vivo (Pochedly, 1977). Interestingly, neuroblastoma cells can either be maintained in vitro as proliferating and undifferentiated but committed, neuroblast-like cells or induced to differentiate after certain modifications of the culture medium. Upon differentiation, neuroblastoma cells become amitotic and extend neurites which show some characteristics of axon-like processes (Seeds et al. 1970; Yamada et al. 1970, 1971; Prasad, 1975).

As initially inferred from pharmacologic and ultrastructural studies (Seeds et al. 1970; Yamada et al. 1970,1971), microtubules are actively involved in neurite growth; however the molecular mechanisms responsible for the regulation of the assembly and organization of microtubules within growing neurites have not yet been defined. However, studies of the in vitro assembly of microtubule protein isolated from brain extracts have provided some clues. It has been shown that the presence of certain tubulin-binding proteins, named microtubule-associated proteins (MAPs), is essential for the efficient assembly of microtubules in brain extracts. The addition of MAPs to pure tubulin solutions lowers the tubulin concentration required for microtubule assembly and stabilizes the resulting microtubules (see reviews by Olmsted, 1986; Matus, 1988; Diaz-Nido et al. 1990a). Furthermore, these properties of MAPs can be modulated by post-translational modifications such as phosphorylation and dephosphorylation (see review by Avila and Diaz-Nido, 1991). Thus, MAPs are the best candidates for the factors responsible for promoting and stabilizing the assembly of microtubules in developing neurites.

The major brain MAPs include MAP1A (350 ×10s Mr), MAP1B which is also referred to as MAP5, MAP1X or 1.2 (325 ×103 MJ, MAP2 (270 ×103 MJ and tau proteins (50 to 70 × 103 MJ (apparent molecular weights determined from electrophoretic mobilities in SDS-containing acrylamide gels). Interestingly, the expression of brain MAPs is developmentally regulated. In situ hybridization studies have demonstrated the abundance of MAP1B mRNA both in proliferating and differentiating cells of the developing central nervous system, whereas MAP2 mRNA is only present in mature neurons (Tucker et al. 1989). Although a minor presence of MAP1A mRNA can be detected in the developing brain (Tucker et al. 1989), MAP1B is much more abundant in embryonic (and newborn) than in adult rat brain, while the converse is found for MAP1A, thus suggesting a specific role for MAP1B in developing neurones (Matus, 1988). As for tau proteins, different isoforms arise from alternatively spliced transcripts originating from a single gene (Goedert and Jakes, 1990; Kosik et al. 1989; Himmler et al. 1989). Whereas the carboxy-terminal domain of tau from developing brain contains three repeated sequences which represent tubulin-binding motifs, some adult tau isoforms contain four tubulin-binding motifs and show therefore an increased affinity for microtubules (Goedert and Jakes, 1990).

Here we have examined the possible roles of these MAPs and their post-translational modifications in neurite growth during neuroblastoma cell differentiation.

Detailed descriptions of experimental procedures have previously been published. Clonal cells lines N2A and NIE-115, derived from the same murine neuroblastoma tumour C-1300 (Seeds et al. 1970; Yamada et al. 1970), were used. Neuroblastoma cell differentiation was induced after the incubation of cells in serum-free culture medium. (Diaz-Nido et al. 1988; Diaz-Nido and Avila, 1989).

Cell culture, labelling and immunofluorescence microscopy were performed as described before (Diaz-Nido et al. 1988; Diaz-Nido and Avila, 1989). Antibodies used in this study include polyclonal and monoclonal antibodies to brain MAP1B (Diaz-Nido and Avila, 1989), a monoclonal antibody that recognizes phosphorylated MAP1B (Diaz-Nido et al. 1991a), a polyclonal antibody to /?3 tubulin (Diaz-Nido et al. 1990c), a polyclonal antibody to brain MAP2 (Valdivia et al. 1982; Hernández et al. 1989) and a polyclonal antibody to brain tau proteins (Montejo et al. 1986).

Protein preparations were as in Diaz-Nido et al. (1988). Microtubule protein from neuroblastoma cells was obtained following either the taxol-dependent procedure of Vallee (1982) or the in situ extraction procedure of Black et al. (1984) with slight modifications (Diaz-Nido et al. 1990c). Tau protein from neuroblastoma cells was obtained following the method of Lindwall and Cole (1984). Immunoprecipitation, gel electrophoresis and immunoblotting were performed as described before (Diaz-Nido and Avila, 1989).

Cell fractionation was performed by low speed centrifugation of neuroblastoma cell homogenates in the presence of 0.25 m sucrose and 0.1% (w/w) triton-X-100 (Birnie, 1978). The resulting supernatant was referred to as the ‘cytoplasmic’ fraction and the resulting pellet was referred to as the ‘crude nuclear’ fraction. Crude nuclear fraction was extracted as in Diaz-Nido and Avila (1989), and the cytoplasmic fraction was processed in the same way. Heparin agarose chromatography was performed as described by Mulner-Lorillon et al. (1988).

In vitro phosphorylation assays were performed as described before (Diaz-Nido et al. 1988; Diaz-Nido et al. 1990c).

Microtubule protein from N2A and NIE-115 cells was prepared according to the method of Vallee (1982) and its composition compared with that of rat brain microtubule protein. As shown in Fig. 1, a much simpler composition is found for neuroblastoma microtubule protein; this preparation contains a major 325 ×lO3Afr MAP in addition to tubulin. Immunoprecipitation and immunoblotting experiments indicate that this major MAP corresponds to brain MAP1B (Diaz-Nido et al. 1988; Diaz-Nido and Avila, 1989). Part of the minor 260 × 103 Mr protein which is observed in neuroblastoma microtubule protein appears to be a proteolytic product derived from MAP1B, as previously described (Gard and Kirschner, 1985; Diaz-Nido et al. 1988; Diaz-Nido and Avila, 1989). Tubulin and the protein identified as MAP1B are also the major components of microtubule preparations obtained by the in situ extraction procedure of Black et al. (1984) as shown before (Diaz-Nido et al. 1990c).

It should be noticed that the tubulin and MAP1B contents of neuroblastoma cells are notably higher than those observed in non-neural cell lines (Diaz-Nido and Avila, 1989). Furthermore, the minor neuronal-specific γ3 tubulin isotype is also present in neuroblastoma cells (Diaz-Nido et al. 1990c). These data indicate that neuroblastoma cells can be actually considered as committed neuroblast-like cells.

To identify other minor MAPs present in neuroblastoma cells, immunoblotting experiments with polyclonal antibodies raised against brain MAPs were performed. Using this approach, we have identified several polypeptides immunologically related to brain tau proteins (see below, Fig. 4). These tau-related proteins are soluble in perchloric acid, which is consistent with the reported solubility of brain tau (Lindwall and Cole, 1984). With respect to MAP2, only a minimal amount of immunoreactive protein has been detected in mouse neuroblastoma cells (data not shown). Thus it appears that MAP1B and tau-related proteins are the main MAPs in mouse neuroblastoma cells which are related to those found in brain.

These results have been confirmed by immunofluorescence analyses. It thus appears that MAP1B (Pig. 2) and tau-related proteins (Fig. 3) are present in neuroblastoma cells, particularly within growing neurites. Neurites also contain the neuronal-specific β 3-tubulin isoform (data not shown). However, MAP2 immunoreactivity, which is quite faint in cell bodies as compared with MAP1B immunoreactivity, is completely absent from neurites (Fig. 2). The absence of MAP2 from neurites suggests that it is not involved in neurite outgrowth. Whether the MAP2-related proteins present in cell bodies are actually bound to microtubules or other cytoskeletal elements (i.e. vimentin intermediate filaments) is not clear at present. However, the distribution of MAP2 is highly similar to that of vimentin filaments (data not shown), which favours the latter possibility. Evidence for the association of MAP2 with intermediate filaments in brain cultured cells has been previously described by Bloom and Vallee (1983).

The growth of neurites during neuroblastoma cell differentiation requires the assembly of their inner cytoskeletal scaffoldings, mainly constituted by microtubules (Seeds et al. 1970; Yamada et al. 1970, 1971; Diaz-Nido et al. 1990a). Thus, a net increase in the amount of assembled microtubule protein which parallels neurite outgrowth has been observed after quantifying the amounts of tubulin present in the detergent-soluble protein fraction and in the detergent-insoluble protein fraction (assembled microtubule fraction) obtained using the in situ extraction procedure of Black et al. (1984). Whereas approximately 18% of total tubulin is found polymerized (detergentinsoluble) in undifferentiated NIE-115 cells, approxi-mately 65 % of total tubulin is found in the assembled pool in differentiated NIE 115 cells (Fig. 4).

However, no significant changes in the total amounts of either tubulin or MAPs are observed during neuroblastoma cell differentiation. Fig. 4 shows that the radioactivity associated with tubulin, MAP1B and tau-related proteins (proteins ofMr × 10−3 50, 55,68 and 100) is similar for undifferentiated and differentiated neuroblastoma cells labeled with [35S]methionine (except for the minor 100 ×103Mr tau component, that increases). Quantitative immunoblotting of cell extracts has confirmed these results (data not shown).

Thus, the increase in the amount of assembled tubulin appears not to be driven by a concomitant increase in the synthesis of either tubulin or major MAPs during neurite outgrowth. Other regulatory factors, possibly post-translational modifications of preexisting proteins, are therefore implicated.

In a pioneering effort to identify whether there were post-translational modifications of microtubule proteins during neurite outgrowth, Gard and Kirschner (1985) described the phosphorylation of a minor β-tubulin isoform and a high molecular weight microtubule-associated protein tentatively identified as MAPI. Eddé et al (1987) identified the phosphorylated β-tubulin isoform as one of the neural-specific isotypes. They found that the expression of this isotype is connected with neuronal commitment, whereas its phosphorylation occurs during neurite outgrowth. We have identified this isotype as β 3 tubulin, the product of the M/36 gene, and determined that the phosphorylation site corresponds to serine 444, which is located at the carboxy terminus of the molecule (Diaz-Nido et al. 1990c). The enzyme responsible for this modification is probably a casein kinase Il-related enzyme (Serrano et al. 1987; Diaz-Nido et al. 1990c).

As for phosphorylated MAPI, it corresponds to MAP1B, as determined by immunoprecipitation and phosphopeptide mapping, and also seems to be phosphorylated by a casein kinase Il-related enzyme (Diaz-Nido et al. 1988).

However, little is known about the possible modifications of other minor MAPs. Interestingly, the phosphorylation of tau-related proteins (particularly of a 65 × 10s Mr polypeptide) is also increased in differentiated neuroblastoma cells (Fig. 5). It has been reported that brain tau is a good substrate for casein kinase II in in vitro phosphorylation assays (Diaz-Nido et al. 1987; Correas et al. 1991). It appears that casein kinase II preferentially phosphorylates sites located on the amino-terminal portion of the tau molecule, which is obtained after NTCB digestion (Correas et al. 1991). The amino-terminal portions of some tau-related proteins seem also to be major sites of phosphorylation in differentiated neuroblastoma cells (data not shown). Whether or not a casein kinase II-related enzyme is responsible for the phosphorylation of the tau-related protein in differentiated neuroblastoma cells remains to be established.

The relationship of these phosphorylation events with neurite extension does not only come from their temporal appearance. Thus, an intense immunostaining for phosphorylated MAP1B is observed within developing neurites (Fig. 6) and phosphorylated MAP1B is associated with the detergent-insoluble (assembled) microtubule pool, as shown before (Diaz-Nido et al. 1990c).

As microtubule protein phosphorylation has been partly ascribed to the increased cytoplasmic concentration of casein kinase II (CK II), based on immunofluorescence analyses (Serrano et al. 1989), we have determined CK II activity from crude nuclear and cytoplasmic fractions of undifferentiated and differentiated neuroblastoma cells. Fig. 7 shows that neuroblastoma cell differentiation is accompanied by a significant reduction in the CK II activity purified from the crude nuclear fraction and an increase in the CK II activity purified from the cytoplasmic fraction.

Several possibilities may account for these results. A translocation of the enzyme from the nucleus to the cytoplasm may occur. A block in the entry of the enzyme into the nucleus may also take place, with the concomitant accumulation of the nascent enzyme within the cytoplasm. Furthermore, an activation of the cytoplasmic enzyme together with an inactivation of the nuclear enzyme may occur in parallel. Whereas immunocytochemical data favour the first two possibilities, the latter cannot be ruled out.

To test for alterations in other protein kinase activities, we have measured the level of protein kinase C (PKC), which has been implicated in neuroblastoma cell differen-tiation (Tsuda et al. 1989; Miñana et al. 1990). PKC activity is significantly lower (60%) in differentiated neuroblastoma cells when compared to undifferentiated neuroblastoma cells. We still do not know whether the changes in CK II and PKC activities are independent of each other or are related. However, we have determined that the addition of H7, a potent PKC inhibitor, to culture medium of neuroblastoma cells does not produce the dramatic changes in protein phosphorylation pattern and the full neurite outgrowth which occur after serum withdrawal. Nevertheless, in the presence of H7, the induction of short cytoplasmic extensions is observed; the average length of these extensions is about four times shorter than that of neurites found upon serum deprivation (data not shown). These results agree well with those previously reported by Miñana et al. (1990).

It appears plausible that protein phosphorylation may provide the trigger that activates the assembly of microtubules and/or their interactions with other cyto-skeletal and membrane organelles leading to neurite outgrowth. A casein kinase Il-like enzyme might be implicated in it, although the involvement of other protein kinases cannot be ruled out.

Nothing is known about the molecular mechanisms controlling protein phosphorylation in differentiating neuroblastoma cells. Neuroblastoma cell differentiation is induced by serum deprivation, which means that, in the presence of serum, there must be neurite outgrowth inhibitory factors. It has been suggested that these factors are serum proteases (Cunninghan and Gurtwitz, 1989; Monard, 1988; Diaz-Nido et al. 19916; Shea et al. 1991). The reduction in extracellular proteolytic activity caused by withdrawal of serum or the addition of protease inhibitors to the culture medium might have an effect in increasing the density and stability of adhesion molecules, thus allowing a stronger cell attachment to the substratum. This might in turn produce the generation of specific intracellular signals. Lander (1990) has reviewed three different mechanisms that could connect cell surface recognition events with cytoplasmic signalling events. First, adhesion molecules could interact directly with signalling enzymes such as G proteins, phospholipases, phosphatases or protein kinases. Second, adhesion molecules could interact with the cytoskeleton in a way that would affect to signalling molecules. Third, stretchsensitive ion channels could transduce mechanical tension into calcium influx. Some evidence may favour the first possibility (Schuch et al. 1989). Thus, a decrease in phosphoinositide turnover, followed by a decrease in PKC activity, an increase in the free cytoplasmic calcium concentration and a decrease in intracellular pH have been observed in response to adhesion molecule binding in some cultured cells (Schuch et al. 1989).

The decreased PKC activity in response to adhesion molecule binding can be correlated with that found in neuroblastoma cells after serum withdrawal. Moreover, the addition of PKC inhibitors to culture medium of mouse neuroblastoma cells induces neurite outgrowth (Tsuda et al. 1989; Miñana et al. 1990). However, neuroblastoma cells differentiated in the presence of PKC inhibitors show shorter neurites than those differentiated in medium deprived of serum, thus suggesting that PKC inhibition alone is not sufficient to elicit full neurite development. Interestingly, the addition of some protease inhibitors to the culture medium of neuroblastoma cells also results in short neurites, whereas other protease inhibitors (or serum withdrawal) lead to long neurites (Diaz-Nido et al. 19916). This suggests the involvement of multiple proteases in the control of neurite outgrowth, as previously suggested (Diaz-Nido et al. 19916; Shea et al. 1991). Perhaps the inhibition of certain proteases could result in PKC inhibition and short neurite growth, whereas the inhibition of other proteases could trigger additional signalling pathways resulting in full neurite development (see Fig. 8 for a model).

The additional signalling events that trigger the activation of cytoplasmic casein kinase II remain to be established. An apparent partial translocation of casein kinase II from the nucleus to the cytoplasm of differentiating neuroblastoma cells has been observed by immunofluorescence microscopy (Serrano et al. 1989). Casein kinase II is a ubiquitous protein kinase that has been found in both particulate and soluble subcellular fractions from most mammalian cells (Hathaway and Traugh, 1982), including neurones (limoto et al. 1989; Girault et al. 1990). A remarkable shift toward an increased nuclear concentration during active proliferation of bovine adrenocortical cells has been described (Filhol et al. 1990). This is consistent with its putative role in the regulation of cell proliferation through the phosphorylation of several nuclear proteins, including some oncoproteins (Krebs et al. 1988; Carroll et al. 1988). The shift toward an augmented cytoplasmic concentration during neuroblastoma cell differentiation could then be connected with its putative role in microtubule protein phosphorylation. Thus it is tempting to speculate that different signal transduction pathways leading to either cell proliferation or cell differentiation could modulate the sorting of casein kinase II between the nucleus and the cytoplasm. Whether this modulation may depend on post-translational modifications of casein kinase II awaits further investigation.

The phosphorylation of different microtubule components in neuroblastoma cells parallels the assembly of dynamic microtubules in the growth cone and their progressive stabilization within the neurite shaft (Avila and Diaz-Nido, 1991; Diaz-Nido et al. 1991a).

However, the functional meaning of microtubule protein phosphorylation events is still unclear. Phosphorylation of MAP1B may be a prior step to the assembly of microtubules at the growth cone, since it precedes the increase in microtubule assembly that accompanies neurite outgrowth (Diaz-Nido et al. 1988) and is not inhibited by microtubule-depolymerizing drugs (Gard and Kirscher, 1985; Diaz-Nido et al. 1988). This view is supported by immunocytochemical studies showing the presence of phosphorylated MAP1B within developing neurites, mainly in growth cones (Diaz-Nido et al. 1991a). We have proposed that phosphorylated MAP1B might promote microtubule nucleation (Diaz-Nido et al. 1988, 1990a,b, 1991a). Supportive of this view is the localization of phosphorylated MAP1B to major sites of microtubule nucleation, including the centrosomes of undifferentiated neuroblastoma cells and non-neuronal cells (Diaz-Nido and Avila, 1989; Diaz-Nido et al. 1991a). Interestingly, either the incubation with a monoclonal antibody to a conserved phosphorylated epitope present in different cytoskeletal proteins, including MAPI (Vandré et al. 1986), or the exhaustive phosphatase treatment of centrosomes in situ, block the nucleation of microtubules from these microtubule-organizing centers (Centonze and Borisy, 1990). This strongly suggests that microtubule nucleation by centrosomes depends on the presence of certain phosphoproteins, possibly including MAPlB-re-lated proteins. However, growing neurites do not contain defined microtubule-organizing centers and it has been hypothetized that a unique class of super-stable microtubules may act as nucleating elements (Baas and Black, 1990). Perhaps the phosphorylation of MAP1B is required for the building of these stable microtubules. Evidence for this could be the reported association of a highly phosphorylated form of MAP1B, referred to as MAP5a, with a stable cytoskeletal fraction obtained from brain homogenates (Riederer et al. 1990).

Phosphorylation of β 3 tubulin probably occurs after microtubule assembly, as the kinetics of phosphorylation are delayed with respect to the augmented microtubule assembly that parallels neurite growth (Diaz-Nido et al. 1990c) and inhibited when neuroblastoma cells are treated with microtubule-depolymerizing drugs (Gard and Kirscher, 1985; Serrano et al. 1987). These data are also consistent with the fact that assembled tubulin is a better substrate for casein kinase II in vitro than unassembled tubulin (Serrano et al. 1987; Diaz-Nido et al. 1990c). We have proposed that assembled tubulin within the neurite shaft could be phosphorylated in order to stabilize neurite microtubules (Diaz-Nido et al. 1990c, 1991a). It can be hypothetized that the phosphorylation of the carboxy terminus of the molecule, which constitutes the MAPbinding domain (Serrano et al. 1984, 1985; Avila, 1991), augments its negative charge, thus favouring its binding to the tubulin-binding motifs of MAPs which are cationic (Avila, 1991; Cross et al. 1991). This would imply a stronger MAP binding to tubulin, and consequently an enhanced microtubule stability within axon-like processes.

The role of tau phosphorylation in neuroblastoma cells is not yet clear. It has been shown that the phosphorylation of brain tau proteins at sites within the carboxyterminal portion of the molecule close to the tubulin-binding motifs may reduce their association with microtubules (Steiner et al. 1990). Nevertheless, the functional consequences of phosphorylation at the amino-terminal portion of the tau molecule are still entirely unknown.

It is worth stressing that phosphorylation events similar to those described in mouse neuroblastoma cells during neurite growth are observed in other neuronal cell types both in culture and in vivo.

The extension of neurites from cultured rat PC 12 cells treated with nerve growth factor (NGF) has been particularly well studied (Greene, 1984). These chromaffin-like cells assume many of the features of sympathetic neurones when cultured in the presence of NGF. Exposure of PC12 cells to NGF causes a slow extension of neurites after a lag of approximately 18 h. Initiation of neurite outgrowth can be suppresed by inhibitors of RNA synthesis. Interestingly, the induction of both MAP1B and tau protein expression is observed during this time interval (Greene et al. 1983; Drubin et al. 1985). The phosphorylation of /3-tubulin, MAP1B and other MAPs referred to as chartins has also been observed (Greene et al. 1983; Aletta and Greene, 1987; Aletta et al. 1988). However, when PC12 cells are ‘primed’, i.e. treated with NGF and then divested of their neurites by NGF deprivation, and subsequently recultured in the presence of NGF, neurites grow quite rapidly in a way which is largely insensitive to inhibitors of RNA synthesis. Under these conditions, the phosphorylation of /3-tubulin, MAP1B and chartins also accompanies the outgrowth of neurites (Aletta et al. 1988). These results strongly suggest a similarity between ‘primed’ PC 12 cells and neuroblastoma cells.

The presence of phosphorylated MAP1B within growing axons has also been confirmed by immunocytochemical studies of primary cultures of embryonic brain neurones (Diaz-Nido and Avila, 1989; Mansfield et al. 1990). Moreover, in situ immunohistochemistry has revealed the preferential localization of phosphorylated MAP IB within developing axons of the central nervous system (Sato-Yoshitake et al. 1989; Schoenfeld et al. 1989). These results are consitent with the observations that MAP1B is the major in vivo labelled phosphoprotein in microtubule preparations obtained from developing rat brain after an intracraneal injection of radioactive phosphate (Diaz-Nido et al. 19905). MAP1B is also a major phosphoprotein in growth cone preparations obtained from developing rat brain (Mansfield et al. 1990).

The presence of in vivo phospholabelled /33 tubulin and tau proteins in microtubule protein preparations obtained from rat brain after intracranial injections of radioactive phosphate has also been reported (Diaz-Nido et al. 19906,c). The in vivo phosphorylated residue on rat brain /33 tubulin has been identified as the same serine 444 at the carboxy terminus of the molecule that is phosphorylated in mouse neuroblastoma cells (Diaz-Nido et al. 1990c). Futher studies are required to localize the phosphorylation sites on tau proteins. On the other hand, no immunocytochemical data are currently available about the distribution of phosphorylated j33 tubulin and tau-related proteins in primary cultures of neurones or in the developing brain. Even with these reservations in mind, it appears highly plausible that phosphorylation events similar to those found in neuroblastoma cells (MAP1B, β 3 tubulin and tau) might be occurring within in vivo developing neurones at the time of axonal growth.

Whether some of these phosphorylation events may also underlie forms of neuronal plasticity, both normal and pathological, during brain development and ageing is now a matter of speculation. For instance, Viereck and Matus (1990) have proposed a correlation between MAP1B phosphorylation and the ability of certain axons to regenerate. Likewise, the localization of hyperphosphorylated MAP1B to the neurofibrillary tangles characteristic of brains from Alzheimer’s disease patients has been hypothetically connected with the massive and aberrant neurite regeneration supposedly associated with this disorder (Hasegawa et al. 1990). Hyperphosphorylation of tau proteins is thought of as a hallmark of dementias of Alzheimer’s type (Bancher et al. 1989; Flament et al. 1990). Although multiple protein kinases are possibly involved in pathological tau hyperphosphorylation (Correas et al. 1991), a role for casein kinase II has also been proposed on the basis of its altered level and distribution pattern in brains of patients with Alzheimer’s disease (limoto et al. 1990).

We thank Drs C. Alberto and F. Moya (Faculty of Medicine, University of Alicante, Spain) for antibody to phosphorylated MAP1B. We also thank Ms Nancita R. Lomax (Drug Synthesis and Chemistry Branch, Division of Cancer Treatment, National Cancer Institute, USA) for generously providing taxol. This work was supported by Spanish CICYT.