Expression of Tau Protein in Non-Neuronal Cells: Microtubule Binding and Stabilization

Journal of Cell Science 102, 227-237 (1992) 227 Printed in Great Britain © The Company of Biologists Limited 1992

Expression of tau protein in non-neuronal cells: microtubule binding and stabilization

GLORIA LEE and SUSAN L. ROOK

Program in Neuroscience, Harvard Medical School and Center for Neurologic Diseases, Department of Medicine (Division of Neurology), Brigham and Women's Hospital, Boston, MA 02115, USA

Summary

The microtubule-associated protein tau is a developmen- tubule contacts or for proper protein folding in the cell. tally regulated family of neuronal phosphoproteins that In our expression system, the bundling of cellular promotes the assembly and stabilization of micro- microtubules occurs only in transfections using four- tubules. The carboxy-terminal half of the protein repeat tau constructs; any four-repeat construct capable contains three copies of an imperfectly repeated se- of binding is also able to induce bundling. Our data quence; this region has been found to bind microtubules suggest that the presence of bundles is correlated with in vitro. In addition, a fourth copy of the repeat has been enhanced microtubule stability; factors that increase found in adult-specific forms of tau protein. To examine stability such as higher levels of tau protein expression or the structure and function of tau protein in vivo, we have the presence of the fourth repeat, increase the fraction of transiently expressed fetal and adult forms of tau protein transfected cells showing bundles. Finally, the presence and tau protein fragments in tissue culture cells. of tau protein in the cell allows all interphase micro- Biochemical analysis reveals full-length products with tubules to become acetylated, a post-translational modi- heterogeneity in post-translational modification syn- fication usually reserved for a subset of stable cellular thesized in the cells. Immunofluorescent staining of microtubules. transfected cells shows that, under our conditions, sequences on both sides of the repeat region are required for in vivo microtubule co-localization. These additional Key words: transient transfections, microtubule bundling, regions may be required either for enhancing micro- microtubule stability, acetylated tubulin.

Introduction The earliest purification of tau protein revealed a family of related phosphoproteins (Cleveland et al., Microtubules are ubiquitous structures in eukaryotic 1977). It is now clear that some of this heterogeneity is cells and perform a variety of cellular functions. The generated by alternative splicing (Himmler, 1989). In ability of microtubules to achieve such functional human, cDNA clones for six tau isoforms and their diversity has been thought to be partially provided developmental expression have been described through the assistance of various associated proteins. (Goedert et al., 1988; Goedert et al., 1989a,b). Human Tau protein is a microtubule-associated protein found adult-specific forms are distinguished by the presence of primarily in neuronal tissue. It promotes microtubule one or more of exons2, 3 or 10 (Goedert et al., 1989a,b; assembly in vitro (reviewed by Olmsted, 1986) and exon number as assigned by Himmler (1989) for bovine stabilizes cellular microtubules when microinjected into gene). The most striking feature of the primary cells (Drubin and Kirschner, 1986). Expression of tau structure of tau protein as predicted from cDNA clones protein has been correlated with the extension of is the presence of three evenly spaced, imperfectly neurites during the differentiation of pheochromo- repeated copies of an 18 amino acid sequence in the cytoma cells (Drubin et al. 1985). Moreover, the carboxy-terminal half of the protein. One or more presence of tau antisense in primary neuronal cell copies of the repeat are able to bind to microtubules in culture appears to block the development and mainten- vitro (Butner and Kirschner, 1991; Himmler et al., ance of axon-like processes, suggesting a role for tau 1989; Lee et al., 1989). Not surprisingly, increasing the protein in the establishment of neuronal cell polarity number of copies of the repeat increased the efficiency (Caceres and Kosik, 1990; Caceres et al., 1991). In situ, of microtubule binding (Butner and Kirschner, 1991; tau protein is associated with axons (Binder et al., 1985; Lee et al., 1989). Synthetic one-repeat peptides also Kowall and Kosik, 1987; Brion et al., 1988; Trojanowski promote microtubule assembly; however, the required et al., 1989). stoichiometry indicates that the peptides are much less 228 G. Lee and S. L. Rook potent than intact protein (Ennulat et al., 1989; Joly et SV40 early promoter, filled in, then ligated to an Ncol linker a!., 1989). The high molecular weight microtubule- in order to introduce an Ncol site at this location. The sequence of the new multiple cloning site in this vector, associated protein MAP2 shares about 50% amino acid + homology with tau protein at the carboxy-terminal end named pECEN , was checked by DNA sequencing. 220 residues and also contains three repeats (Lewis et Most three-repeat human tau inserts were prepared by polymerase chain reaction (PCR) using pl9tau plasmid DNA al., 1988). As expected, a fragment of MAP2 from this (Lee et al., 1989) as template. PCR primers contained area coassembles with microtubules in vitro (Lewis et restriction sites that enabled insertion into pECEN+. pEn was al., 1988). prepared by cutting and recircularizing pEnl23c. In some The adult-specific exon 10 encodes a 31 amino acid plasmids, translation was terminated by a stop codon in the sequence that contains a fourth copy of the microtubule primer; in other plasmids, translation was terminated by the binding repeat unit. The inclusion of this fourth repeat stop codons in the vector. Table 1 lists all plasmids used in this increases the activity of tau protein in in vitro study; tau sequences expressed and new sequences contribu- microtubule assembly assays (Goedert and Jakes, ted by primers, linkers or vectors are shown. The longest 1990). The expression of a four-repeat tau protein in sequence contributed in this manner was 11 residues and its presence did not correlate with any functional property of the vivo results in the bundling of cellular microtubules expressed proteins. (Kanai et al., 1989). pEnl234c was constructed from pEnl23c by replacing the In this study, we examined the in vivo structure and tau cDNA carboxy-terminal 500 bp with the analogous 593 bp function of fetal and adult tau protein. Transiently fragment from an adult human tau four-repeat cDNA clone expressed tau protein and tau protein fragments were (Mori et al., 1989) kindly provided by Dr. H. Mori (Tokyo characterized using Western blot and immunoprecipi- Metropolitan Institute of Gerontology, Tokyo, Japan). The tation protocols. The ability of the expressed protein to resultant cDNA is a full-length four-repeat tau cDNA associate with cellular microtubules and to induce identical to a clone previously identified by Goedert et al. microtubule bundling were determined by immunoflu- (1989a). pE1234c(487) and pE1234(487) were constructed orescence. Our results indicate that, under our con- from pE123c(487) by similar replacements. pEnl23(4c) was ditions, sequences adequate for in vitro binding are not prepared by cutting and recircularizing pEnl234c. The remaining human four-repeat tau inserts were prepared by adequate for in vivo co-localization to microtubules. PCR from pEnl234c. Polymerase chain reaction primers used Furthermore, the presence of adult-specific tau protein to prepare inserts for insertion into pECEN+ were identical to can induce microtubule bundling; microtubule stability those used for the three-repeat human tau inserts and, is also increased. Bundling and stability may be related, therefore, the starts and stops of expressed sequences are since no evidence for cross-linking of microtubules by identical. tau protein was found. Lastly, the presence of tau Because most constructs were made using polymerase chain protein correlated with the acetylation of all cellular reaction technology, we acknowledge the small probability of microtubules, indicating enhanced stability for all a base change occurring during synthesis. However, most of microtubules. the DNA fragments prepared were fairly small and the biochemical characterizaton of expressed protein (see below) would detect the inadvertant production of a termination codon. Also, all of the constructs overlapped, so that if a single amino acid replacement were to alter dramatically the Materials and methods properties of the expressed protein, an inconsistency would be noted. Construction of expression plasmids The eukaryotic expression vector pECE (Ellis et al., 1986) Transfections was modified in the following manner to introduce a unique CHO or 3T3 (NIH) cells were grown in oMEM sup- Ncol cloning site into its multiple cloning site. First, the sole plemented with 10% fetal bovine serum (Hyclone, Logan, Ncol site in pECE was removed by Ncol cleavage followed by UT) and seeded onto 12 mm coverslips in 24-well plates at a fill-in and re-circularization. The plasmid was then cut with density of approximately 0.2 x 104 /well. CHO cells were Bglll, the first site in the polylinker site downstream from the transfected by calcium phosphate precipitation using standard

Table 1. Eukaryotic expression plasmids

NH2-terminal sequences Tau sequence COOH-terminal sequences Plasmids from primers, linkers or vector (residue no.) from primers, linkers or vector pEnl23c/pEnl234c - 1-352/383 * pEn - 1-163 GIRALDK* pEnl - 1-224 GEFEL* pEnl23(921)/pEnl234(921) - 1-307/338 TVPRG1RALDK* pEnl23(843)/pEnl234(843) - 1-281/312 * pEnl23(4c) — 1-285 GNSSSR* pE123(517)/pE1234(517) MG 173-307/538 TVPRG1RALDK* pE123(487)/pE1234(487) MAG 164-307/338 TVPRGIRALDK* pE l23c(487)/pE1234c(487) MAG 164-352/383 * pE123c(517)/pE 1234c(517) MG 173-352/383 *

•Indicates STOP codon. Expression of tau protein in 3T3 cells 229 techniques (Ausubel et al., 1989). 3T3 cells were transfected Western blotting and immunoprecipitation by calcium phosphate precipitation as modified by Chen and Cell lysates for Western blot analysis were prepared from Okayama (1988). A2ftg sample of plasmid DNA was added transfected 3T3 cells by using 60 y\ boiling of Laemmli SDS per well. sample buffer (Laemmli, 1970) to scrape transfected cells Nocodazole treatment of the transfected cells was per- from 24-well plates. Lysates were electrophoresed on 7.5% to formed by adding nocodazole (Aldrich Chemical Co., 15% gradient gels, then transferred to Immobilon (Millipore Milwaukee, WI) to a final concentration of 3.3 /iM, then Corp., Bedford, MA). The blot was probed with affinity- incubating for the times indicated. Triton extractions were purified antibody to tau protein and I-labeled goat anti- performed by first washing cells with extraction buffer minus rabbit secondary antibody (ICN Radiochemicals, Irvine, Triton at 37°C, then incubating with extraction buffer (80 mM CA). Pipes, pH 6.8, 1 mM MgCI2, 1 mM EGTA, 0.1% Triton- An estimation of the average amount of tau protein X100, 30% glycerol) at 37°C (Solomon et al., 1979). synthesized per transfected cell was made on the basis of the Extraction buffer was then removed and cells washed with transfection efficiency and the amount of tau protein detected extraction buffer minus Triton, then with PBS (137 mM NaCI, by Western blotting. The quantity of tau protein present in a 4 2.7 mM KC1, 10 mM PO4, pH 7.4) before fixation. The time harvested cell lysate (2X10 cells) was calculated by excising of extraction had been determined by inspecting the cells after the Immobilon band corresponding to the expressed protein, extraction using anti-tubulin staining and by immunoblotting counting the 125I-labeled goat anti-rabbit secondary antibody Triton-soluble and -insoluble fractions with anti-tubulin. Over bound, and converting the cts/min to protein, using Immobi- extraction (1 min for 3T3 or 2 min for CHO cells) resulted in a lon bands from adjacent lanes processed in parallel, which large increase in Triton-soluble tubulin as detected by contained known amounts of fetal brain tau protein or immunoblot and ruptured cytoskeletal network as detected by Escherichia coli tau protein. The average amount synthesized immunofluorescence. Therefore, extraction times of 0.5 min per transfected cell equaled quantity in lysate divided by and 1 min were chosen for 3T3 and CHO, respectively. Cells number of transfected cells (2X104 x transfection efficiency; were judged to be fully extracted, since non-cytoskeletal tau transfection efficiency was determined by a parallel transfec- fragments were no longer detectable by immunofluorescence. tion processed for immunofluorescence). Because the tau polyclonal antibody used to probe the blot was derived against Immunofluorescence brain tau protein, the reactivity of the antibody to the Cells were fixed with 0.3% glutaraldehyde in 80 mM Pipes, standard brain tau protein vas greater than the reactivity to E. co/i-synthesized tau protein, which lacks phosphorylated pH 6.8, 1 mM MgCl2, 5 mM EGTA, 0.5% NP40 for 10 minutes at room temperature (Drubin and Kirschner, 1986). epitopes (e.g. 1 ng E. coli protein corresponded to 420 cts/min After washing with PBS, cells were incubated with 10 mg/ml while 1 ng brain protein corresponded to 1516 cts/min). Since it is not known whether the reactivity of 3T3 expressed tau NaBH4 for 7 minutes, washed with PBS, then incubated with 0.1 M glycine for 20 minutes. After washing with PBS, cells protein more closely resembles that of brain protein or E. coli were incubated with methanol for 10 minutes. The initial step protein, the amount of protein expressed is given as a range, of this protocol combines permeabilization and fixation, using E. coli standard for the higher estimate, brain protein because performing permeabilization first results in the for the lower estimate. While in theory, a monoclonal extraction of tau protein from the detergent-insoluble cyto- antibody would have reacted to tau from all sources equally, skeleton. Since full-length tau protein remains bound to in practice this did not work, since the monoclonal we tried microtubules under this fixation protocol, we chose to was unable to yield a signal above background in detecting tau compare all constructs using these identical conditions. Fixed in transfected cell lysates. A comparison of the reactivity of cells were washed with PBS, 10 mg/ml BSA, 0.1% Tween 20 the two antibodies to various quantities of E. coli tau protein prior to antibody staining. showed the signal from the tau monoclonal (5E2) to be 5- to 10-fold less than that from the affinity-purified tau polyclonal The antibody staining protocol entailed staining with the antibody. first primary antibody, washing with PBS, 10 mg/ml BSA, 0.1% Tween 20 (5 X 2 min), staining with second primary For two-dimensional gel analysis, transfections were per- antibody, washing similarly, then staining with both labeled formed in 6-well plates using 5 ng DNA per well. Cells were secondary antibodies. The primary antibodies used were: harvested by scraping with 80 jj\ boiling lysis buffer (0.5% affinity-purified polyclonal antibody against tau protein SDS, 50 mM Tris-HCl, pH 7.6,150 mM NaCI). Samples were (Pfeffer et al., 1983) at 1:1000 for 15 minutes at room electrophoresed on NEPHGE gels (O'Farrell et al., 1977) temperature, monoclonal antibody DMa-1 against tubulin using pH 3-10 Ampholines. The second dimension was 10% (ICN Biochemicals, Inc., Irvine, CA) at 1:500 for 15 minutes SDS/polyacrylamide gel, which was blotted and probed as at room temperature, and monoclonal antibody 6-11B-1 described above. E. coli lysate containing full-length tau against acetylated tubulin kindly provided by Dr. G. Piperno protein was prepared from E. coli BL21(DE3)LysS cells and used at 1:10 overnight at 4°C. Secondary antibodies were transformed with pET-3d expression plasmid (Studier et al., goat anti-rabbit IgG, rhodamine- or fluorescein-labeled 1990) harboring full-length human three-repeat tau cDNA. (Boehringer Mannheim Corp., Indianapolis, IN), and goat Cell lysates for immunoprecipitation were prepared from anti-mouse IgG, rhodamine- or fluorescein-labeled (Boehr- 3SS-methionine-labeled transfected 3T3 cells by using 80 fj\ inger Mannheim Corp., Indianapolis, IN). Secondary anti- lysis buffer supplemented with 2 mM EDTA, 2 mM PMSF bodies were incubated with cells for 30 minutes at room (phenylmethylsulfoxyl fluoride) to scrape transfected cells temperature. Coverslips were mounted in 1 mg/ml p- from 6-well plates. (Each well had received approximately 175 phenylenediamine, 90% glycerol (Johnson and Araujo, jtCi of 35S-methionine for 4 hours prior to harvesting. 1981). Cells were photographed using x63 Neofluar lens on a Harvesting was performed 48 hours after DNA addition.) Zeiss Axioskop. Lysates were boiled and then immunoprecipitated as de- Bundling efficiency was determined by scoring 100-400 scribed by Drubin et al. (1988) with the following modifi- transfected cells for bundles. A cell was scored positive for cations: lysate was precleared twice with Pansorbin and once bundling if one or more obviously thick, dense bands of with Protein A/Sepharose beads (Pharmacia, Piscataway, microtubules were observed in the cell (see Fig. 5, below). NJ), Protein A/Sepharose beads were used for adsorption of 230 G. Lee and S. L. Rook antibody-antigen complexes, and beads were washed 5 times length products are obtained from pEnl23c, in NET (0.5% NP40, 150 mM NaCl, 50 mM Tris-HCl, 5 mM pEnl23(921), pEnl23(843), pEn, PEnl234c, EDTA) buffer prior to elution. Complexes were electrophor- pEnl234(921) and pEnl234(843) (refer to Fig. 1 for tau esed on 15% to 25% or 10% to 20% gradient gels. Gels were sequences expressed). The estimated molecular masses fluorographed prior to autoradiography. were consistently higher than those calculated from primary structure data; this discrepancy may be due in part to phosphorylation of the protein (Lindwall and Results Cole, 1984) and has been noted for tau protein purified from brain as well. While multiple bands were seen in Full-length tau protein and tau protein fragments several lanes, it is unlikely that this reflected degra- expressed in vivo dation, since there were no bands in common among To study the in vivo structure and function of human the constructs, all of which overlapped in sequence. tau protein, a panel of eukaryotic expression plasmids To confirm the lack of degradation for the smaller containing tau sequences was derived from both three- constructs from the carboxy-terminal end, it was repeat and four-repeat tau cDNA clones (Fig. 1A,B). necessary to label the cells with -^S-methionine and The four-repeat tau cDNA used was identical to the immunoprecipitate tau protein as Western blotting was three-repeat clone except for the inclusion of tau gene not sufficiently sensitive to visualize these proteins. Fig. exon 10, which encodes a fourth copy of the repeat. 3 shows the labeled products immunoprecipitated from Most of the deletion constructs were made in the same transiently transfected cells. Plasmids pE123c(487), manner for both the three-repeat and four-repeat pE123c(517), pE123(517), pE1234c(487) and constructs so that the starting and ending sequences of pE1234c(517) exhibited protein patterns that suggest the expressed proteins were identical (Table 1). These that intact protein is synthesized. The molecular masses plasmids were used to express transiently tau protein in calculated from size standards on the gel were much 3T3 cells, which normally do not express tau protein. higher than those calculated from primary sequence To verify the presence of tau proteins in 3T3 cells, the data. For example, the two proteins in pE123c(487) transiently transfected cells were analyzed by Western (Fig. 3A, lane 2) average 32.5 kDa while the primary blotting. Fig. 2 shows that detectable levels of full- sequence calculation yields 20 kDa for residues 164 to 352. Therefore, besides indicating that intact protein is Mfcrotubule bindlnj pEn123c Fig. 1. (A) Schematic diagram of expression plasmid inserts from three-repeat tau cDNA. Each plasmid is pEn constructed in expression vector pECEN+ as described in pEn1 Materials and methods, and is predicted to express the pEn123(843) indicated amino acids from tau protein. The human full- pEn123(921) — 307 length three-repeat clone used was first reported by pE123c(487) Goedert et al. (1988). The three 18 amino acid repeats 173 — identified in mouse tau protein (Lee et al., 1988) are pE123c<517) located at residues 198-215, 229-246 and 261-278 in human; pE123(517) 173 307 each repeat is marked with a short line. Microtubule binding capability, detected by immunofluorescence, is B Wcrotubule shown at the right. pE123(487), which encodes residues Binding BundUng 164-307, was not included because no transfected cells were pEn1234c detected. (B) Expression plasmid inserts from four-repeat — — — J3» tau cDNA. Plasmids were constructed as described above. pEn1234(92l) 311 The human full-length four-repeat tau cDNA used is pEn1234(843) identical to the three-repeat cDNA above, except that it pEM23(4c) 2tS includes exon 10 (Himmler, 1989), which results in the pE1234c<487) insertion of 31 residues at residue 217 of the three-repeat pE1234c<517) 173 — — 3J3 sequence. The inserted sequence is shown by the heavy bar 173 331 pE1234(517) and contains the fourth repeat. Therefore, amino acids 248- M 331 383 of the four-repeat protein correspond to the three- pE1234(487) repeat protein residues 217-352. The repeats are located at residues 198-215, 229-246, 260-277 and 292-309 and are marked with short lines. Microtubule binding and bundling capability for each construct are shown at the right. Relative bundling efficiencies are discussed in Results. (C) Diagram showing the role of various tau protein regions towards in vivo colocalization to microtubules. Numbers denote amino acid residues of three-repeat tau \ \ protein; numbers in parenthesis denote analogous residue 281 307 352 numbers in four-repeat tau protein located downstream (312) (338) (383) from exon 10, which inserts at residue 217. R indicates repeat region with the first repeat starting at residue 198 | required by required only by and the last repeat ending at 281(312). I all isoforms three repeat • not required Expression of tau protein in 3T3 celb 231

CM r-- T- CO CO C\J B 00 CO r- o n N s N N N N B CO CO O co co co w m o o 00 co co co co 8 DC *d" CO xf CO On ffl N (O N

— 43 • —«3

V "-43 -43

-29 "-29 -28 "-29

-18 -14

—*'«•

12 3 4 5 1 Fig. 2. Immunoblot of transiently transfected 3T3 cells. Cells were transfected and harvested as described in Fig. 3. Immunoprecipitation of 35S-labeled tau protein Materials and methods. Lysates were electrophoresed on a fragments from transiently transfected 3T3 cells. Immune 7.5% to 15% gradient gel, blotted and probed with 125 complexes from 3T3 transfected cell lysates were prepared affinity-purified anti-tau followed by I-labeled goat anti- as described in Materials and methods. (A) A 15% to 25% rabbit secondary. (A) Duplicate transfections using gradient gel displaying products from transfections with pEnl23c (1-352), pEnl23(921) (1-307), PEnl23(843) (1-281) three-repeat constructs. Lane 1 shows control and pEn (1-163). Numbers at top indicate three-repeat tau immunoprecipitation from non-transfected cells. Lane 2 residues expressed. (B) Transfections using pEnl234c (1- contains immunoprecipitated products from pE123c(487) 383), pEnl234(921) (1-338) and pEnl234(843) (1-312). transfection expressing residues 164-352, lane 3 Numbers at top indicate four-repeat tau residues pE123c(517) expressing residues 173-352, lane 4 expressed. Molecular mass markers (in kDa) are as pE123(487), and lane 5 pE123(517) expressing residues 173- indicated. 307. Positions of bands marked with arrows (lane 2) correspond to molecular masses of 34 kDa and 31 kDa; bands marked with asterisks (lane 3) 26.3 kDa and 23.8 present, these data also suggest that post-translational kDa; dark band in lane 5, 16 kDa. Molecular mass modification in the 3T3 cells causes dramatic shifts in calculated from primary structure for 164-352 is 20 kDa, the gel mobility of these carboxy-terminal fragments. It 173-352 19 kDa, and 173-307 14.4 kDa. (B) A 10% to 20% is likely that these shifts are caused by phosphorylation; gradient gel displaying products from transfections with it has been suggested that phosphorylaton may change four-repeat constructs. Lane 1 shows immunoprecipitated products from pE1234c(487) transfection expressing the conformation of the molecule such that SDS residues 164-383, lane 2 pE1234c(517) expressing residues binding is increased (Lindwall and Cole, 1984). Two- 173-383, and lane 3 non-transfected control. Positions of dimensional gel analysis shows in vivo expressed tau bands marked with arrows correspond to molecular masses protein to be much more acidic than the identical of 35 kDa and 33 kDa. Molecular masses calculated from protein synthesized in E. coli (data not shown). primary sequence data for residues 164-383 is 23.3 kDa and In Fig. 3A, cells transfected with pE123c(517) (lane for 173-383 is 22.3 kDa. Molecular mass standards are as 3) express two forms that appear to differ in molecular shown. mass by approximately 2.5 kDa. When the carboxy- terminal 45 amino acids was truncated (pE123[517], lane 5), heterogeneity was lost, since only one form was difference. Since the 164-352 products (lane 2) are detected. This suggests that residues 308-352 contain a estimated to be 34 kDa and 31 kDa while the 173-352 site that creates heterogeneity in the gel mobility of the products (lane 3) are estimated to be 26.3 kDa and 23.8 protein. Indeed, Steiner et al. (1990) have reported that kDa, the inclusion of residues 164-172 in these three- phosphorylation of serine residue 327 of tau protein repeat fragments results in an apparent increase in results in a shift in the electrophoretic mobility of tau molecular mass of at least 7 kDa. This implies that protein. However, other sites with this property exist, residues 164-172 may affect the conformation of the since doublet bands are also exhibited by constructs protein, causing the retardation of the fragments in gel lacking this same region (e.g. pEnl23[921], electrophoresis. While the sequence of this region, pEnl234[921]). EPKKVAWR, does not contain any serines or The difference in gel mobility between three-repeat threonines, residue 173 is a threonine, which raises the fragments 164-352 and 173-352 (Fig. 3A, lanes 2,3) is possibility that phosphorylation of this residue could larger than one would expect from a nine amino acid affect gel mobility or SDS binding (the loss of residues 232 G. Lee and S. L. Rook

164-172 could affect the recognition of residue 173 by kinases). The contribution of residues 164-172 towards causing gel mobility shifts is diminished, however, when four repeats are present. Fig. 3B shows that four-repeat fragments 164-383 (lane 1) and 173-383 (lane 2) look identical. We were unable to immunoprecipitate products for pE123(487) (Fig. 3A, lane 4), pE1234(487) and pE1234(517) (data not shown). Since these plasmids had transfection efficiencies of 0-1% as assessed by immunofluorescence (see below), we judged that very little protein was present. Plasmids whose products were detectable by Western blotting had transfection efficiencies of 20-40% while those detectable only by immunoprecipitation had efficiencies of 5-10%. Since the necessity to immunoprecipitate correlated with plasmids encoding carboxy-terminal fragments, we surmised that these protein fragments or their mRNAs were susceptible to cellular degradation. An estimate of the amount of full-length tau protein synthesized in transfected 3T3 cells was made by comparing the Western blot signal to the signal given by known quantities of tau protein (see Materials and methods). By dividing this number by the number of transfected cells in the culture, as determined in a parallel transfection, we estimate that an average of 0.5-2 picograms of full-length tau protein is synthesized in a transiently transfected cell. Requirements for co-localization of tau protein and microtubules in vivo The in vivo function of tau protein and tau protein fragments was assessed by examining transiently transfected CHO or 3T3 cells by immunofluorescence. The cells were fixed 24-48 hours after DNA addition and Fig. 4. Immunofluorescent staining of CHO cells stained with antibodies to tau protein and tubulin, using transfected with tau-expressing plasmids. Cells were double immunofluorescence to show the location of tau transfected, fixed and stained as described in Materials and protein relative to the microtubules. Because the cells methods. (a,b,c) Anti-tau staining. The transfecting transiently express tau protein, the level of expression plasmids were: a, pEnl23c encoding amino acids 1-352; varies from cell to cell, depending on the number of b, pEn encoding amino acids 1-163; and c, pE123 plasmid gene copies present. In looking at many cells, we found encoding amino acids 173-307. (d,e,f) Corresponding anti- tubulin staining for each transfection. Cells in a and d were that only the intensity of the pattern of staining varied Triton-extracted prior to staining as described in Materials with the level of expression; the pattern did not vary. A and methods. Bar, 10 /.an. fraction of the transfected cells had no discernible staining pattern, perhaps due to the cell-cycle stage, and were not included in the analysis. ponents from the cytoplasm, leaving intact micro- Fig. 4 shows tau protein and tubulin staining of CHO tubules unobscured by cytoplasmic staining (Solomon cells transfected with plasmids pEnl23c, pEn and et al., 1979). This procedure did not reveal any fraction pE123. As expected, full-length tau (pEnl23c) co- of cytoplasmically localized tau fragments binding to localizes with microtubules (Fig. 4a) while the amino- microtubules. Also, none of the constructs exhibiting terminal fragment having no repeats (pEn) shows microtubule localization showed significant loss of cytoplasmic rather than cytoskeletal staining (Fig. 4b). binding after Triton extraction. Surprisingly, cells transfected with pE123 did not show Fig. 1 shows the microtubule localization results for tau protein co-localizing to microtubules (Fig. 4c); this our panel of constructs. From our data, we conclude fragment contains the entire repeat region and was that under our conditions of fixation, residues 164 to 307 shown to bind to taxol-stabilized microtubules in vitro of the three-repeat tau protein are required for (Lee et al., 1989). Interaction with the cytoskeleton was microtubule association in the cell. For four-repeat tau confirmed by extraction with 0.1% Triton prior to constructs, deleting residues upstream from residue fixation and immunofluorescence (conditions for ex- 338, which corresponds to three-repeat residue 307, traction were derived as described in Materials and also resulted in loss of binding. However, the required methods). Triton extraction removes soluble corn- sequence at the amino-terminal side of the four-repeat Expression of tau protein in 3T3 cells 233 region was different from that required for the three- repeat protein - a four-repeat construct that was deleted up to 173 still bound. Fig. 1C summarizes the tau protein structural requirements for in vivo microtubule co-localization. The necessity for the extra nine residues in three-repeat constructs may be related to phosphorylation and/or protein folding, as suggested by the immunoprecipitation results. While four-repeat fragments 164-383 and 173-383 co-migrate on gels (Fig. 3B, lanes 1,2) and co-localize to microtubules in vivo, three-repeat fragments 164-352 and 173-352 differ both in their gel migration patterns (Fig. 3A, lanes 2,3) and in their microtubule localization abilities. Fragment 173-352 has increased gel mobility relative to 164-352 and is unable to co-localize with microtubules. Lastly, although our deletion analysis of three- and four-repeat constructs identifies the unique sequence required in vivo as amino acids 164-307 for the three- Fig. 5. Immunofluorescent repeat and 173-338 for the four-repeat protein, con- staining of 3T3 cells structs expressing just residues 173-338 or 164-338 did transfected with four-repeat not bind microtubules in vivo. However, since we could tau cDNA showing microtubule bundling. Cells not detect intact protein by immunoprecipitation for were transfected, fixed and these two plasmids (pE1234[517] and pE1234[487]), we stained as described in cannot rule out the possibility that their inability to bind Materials and methods. The was due to degradation. Alternatively, these fragments transfecting plasmids were: may not have been able to fold properly and required (a.b) pEnl234(921); and (c) inclusion of neighboring regions to achieve functional pE1234c(517). Anti-tau conformation. The localization of the three-repeat staining is shown. Anti- fragment 164-307 could not be determined because no tubulin staining pattern of transfected cells were detected by immunofluorescence transfected cells was virtually indistinguishable from the for this plasmid. Immunoprecipitation also failed to anti-tau pattern. Bar, 10 /m\. detect protein (Fig. 3A, lane 4).

Microtubule bundling boxy-terminal 45 residues and the amino-terminal 172 Cells with full-length four-repeat tau protein residues of tau protein are not required for this result. (pEnl234c) exhibited bundled microtubules similar to those reported in mouse L-cells by Kanai et al. (1989), Tau protein stabilized microtubules in cells where a rat full-length four-repeat tau cDNA was used. Drubin and Kirschner (1986) have shown that tau Bundles were most dramatically seen as sharp spikes protein, when microinjected into cells, stabilizes micro- emanating from the cell or as thick dense bands running tubules against drug-induced depolymerization. We along the edge of cells (Fig. 5). Kanai et al. (1989) have found that cells transfected with four-repeat tau cDNA defined bundles as microtubule structures with 10-100 were able to withstand more prolonged drug treatment times the thickness of normal microtubules. In non- than those transfected with three-repeat cDNA. Micro- transfected 3T3 cells, microtubules did not coalesce into tubule fragments were still visible in cells expressing such structures. In contrast to the cytoplasmic versus four-repeat protein after 1 hour of incubation in 3.3 /xM cytoskeletal localizations described above, the occur- nocodazole (Fig. 6a) while none were present in the rence of bundles appeared to be correlated with the three-repeat protein transfectants (Fig. 6b) and non- level of expression of the four-repeat protein; bundles transfected cells (Fig. 6c). After 3 hours of drug were never found in cells with faint tau staining. treatment, microtubules were absent in all cells. Transfections with four-repeat protein showed an The enhanced stability of tau-associated micro- average of 20% transfected cells containing bundled tubules was further demonstrated by staining trans- microtubules. In contrast, while the three-repeat con- fected cells with an antibody to acetylated tubulin struct consistently had a higher transfection efficiency (Piperno and Fuller, 1985). Microtubules containing than the four-repeat construct, only 1% of cells acetylated-tubulin subunits have been identified as a transfected with the three-repeat protein showed stable subset of microtubules that persist after low-level bundling. When non-transfected cells were scored for nocodazole treatment (Piperno et al., 1987) and turn bundled microtubules, less than 0.25% were positive. over less rapidly in the cell relative to dynamic Using the collection of human tau constructs bearing microtubules (Schulze et al., 1987; Webster and Borisy, four repeats, we found that every tau protein fragment 1989). The top cell in Fig. 7b shows a typical acetylated capable of binding to microtubules also exhibited tubulin pattern in a non-transfected cell. Looking at bundled microtubules (Fig. IB). Therefore, the car- transfected cells that exhibited an interphase micro- 234 G. Lee and S. L. Rook

Fig. 6. Microtubules stabilized by the presence of tau protein against depolymerization by nocodazole. Transfected 3T3 cells were treated with 3.3 fjM nocodazole for 1 hour prior to Triton extraction and fixation as described in Materials and methods, (a) Full-length four-repeat tau-transfected cell; (b) full-length three-repeat tau-transfected cell; and (c) non- transfected cell. The cells shown are stained with anti-tubulin. Transfected cells nocodazole-treated and Triton-extracted were readily distinguishable from non-transfected cells due to background material, presumably free tau and tubulin, located in and around the transfected cells that stained with tau and tubulin antibody. Bar, 10 fan.

linked to its acquisition of acetylated subunits, the presence of tau in the cell does enable a larger set of microtubules to become acetylated than would have otherwise.

Discussion Regions flanking the repeat region contribute to microtubule co-localization in vivo The structure and function of tau protein and MAP2 protein have been studied in vitro using protein fragments synthesized in E. coli or in rabbit reticulocyte lysate systems (Butner and Kirschner, 1991; Himmler et al., 1989; Lewis et al., 1988; Lee et al. 1989). These data suggest that sequences up to residue 190 on the amino- terminal side and sequences downstream from residue 281 on the carboxy-terminal side could be deleted without abolishing in vitro microtubule binding activity. However, our in vivo experiments indicate that additional sequences on each side of the repeat region are necessary for co-localization of tau protein to cellular microtubules. Butner and Kirschner (1991) also found in their in vitro binding study that sequences outside of the repeats (residues 92-190 or 289-352, numbered as in three-repeat tau) can increase the affinity of a tau fragment for microtubules 10- to 20-fold. With MAP2, Fig. 7. Immunofluorescent staining of acetylated Lewis et al. (1989) also found that including residues microtubules in transfected cells. Transfected 3T3 cells 145-193 or 292-352 qualitatively enhanced in vivo were Triton-extracted before fixation and stained with anti- tau (a and c) and anti-acetylated tubulin (b and d) as microtubule binding. Therefore, while the presence of described in Materials and methods, (a and b) A non- three- to four-repeat motifs can result in 80-100% of the transfected cell (top) and a cell expressing three-repeat tau input tau fragment binding to microtubules in vitro (bottom), (c and d) A cell expressing four-repeat tau. Bar, (Butner and Kirschner, 1991), contributions made by 10 ^m. the flanking regions are important for in vivo microtubule co-localization. Our data more precisely define tubule array, many contained an acetylated tubulin the areas of tau protein that enhance microtubule patterni that was virtually identical to the tau-stained interaction. As summarized in Fig. 1C, we find that microtubule pattern (Fig. 7). This occurred using both including up to residue 164 on the amino side or down the three- and four-repeat proteins. Other transfected to residue 307 on the carboxy side of the repeats is cells had the typical acetylated tubulin patterns found in sufficient to differentiate between in vivo binding and non-transfected cells (data not shown). Therefore, not binding in our system. while the presence of tau on a microtubule is not tightly The role of the contributions made by the flanking Expression of tau protein in 3T3 celb 235 regions may be to enhance microtubule interaction by supporting the notion that the carboxy-terminal end of providing additional contacts that by themselves are tau contains a short hydrophobic "zipper" sequence insufficient to bind microtubules. These contacts may that mediates microtubule bundling (Lewis et al. 1989). not be needed for in vitro binding because in vitro assays contain a vast excess of stabilized microtubules, Microtubule stabilization and bundling which may drive the binding reaction. Our in vivo data The deletion analysis of the four-repeat tau protein reflect the cellular conditions normally encountered by indicates that sequences upstream from amino acid 173 tau protein. Besides smaller amounts of microtubules, and downstream from 338 are not required for there may also be the presence of competing endogen- microtubule bundling. We propose that bundling is very ous microtubule binding proteins. Alternatively, the likely the result of enhanced stabilization of micro- flanking sequences may be necessary for efficient tubules, on the basis of the following reasons: (1) the folding of tau fragments. While this may not have been areas of tau protein required for bundling correlate with critical for an in vitro assay with purified components, in those involved in microtubule binding. (2) Bundled vivo, an improperly folded protein may bind elsewhere microtubules are seen mainly in cells with a high level of or lack the higher binding affinity required to yield tau expression. (3) The presence of the required fourth detectable localization to microtubules. While our repeat increases the stability of the microtubules, as fixation conditions may reflect a 'stringent' binding shown by nocodazole resistance. The number of cells requirement, the existence of these non-repeat regions showing bundled microtubules in a given transfection that enhance binding may suggest the existence of would be related to the association constant of the specific functions of tau that require a higher binding stabilizing protein and the level of expression. It is affinity than that solely necessary to bind to micro- probable that, for the induction of bundles, a tubules in vitro. heightened level of expression may compensate for a lower association constant and vice versa. Kanai et al. Microtubule bundling in the presence of tau and (1989) have also suggested that bundling is increased in MAP2 cells with a high level of tau protein expression, on the Our results show that the expression of the fourth- basis of a comparison of transient and stable expression repeat leads to a discernible functional difference in the data. In our study, the constructs with the highest microtubule organization and stability in cells. In our bundling efficiencies were pEnl234c and pEnl234(921). expression system, every four-repeat microtubule- Among the constructs capable of inducing bundles, binding tau construct was more effective in inducing these two also exhibited the highest transfection microtubule bundling than the analogous three-repeat efficiency and expressed picogram levels of protein in construct. The immunofluorescent staining of cells cells, allowing for detection by Western blotting. transiently transfected with a fetal mouse tau ex- Lastly, the propensity for microtubule bundles to form pression plasmid, as reported by Lewis et al. (1989), is in the cell may also depend on the cell type employed, consistent with our results in that very little bundling is since bundling may require re-modeling of existing exhibited by three-repeat forms. However, they found cytoskeleton; the ease with which the existing micro- that MAP2, which also contains three repeats, induces tubules are re-arranged may vary from cell to cell type. bundling. Since MAP2 and tau protein share only 50% The expression of tau protein in baculovirus-infected homology in the carboxy-terminal 220 residues, it is Sf9 cells provides an example of an extremely high level likely that the sequence differences account for the of tau expression (greater than 10-fold of that of difference in bundling capability. Moreover, MAP2 is transfected cells) achieved in the presence of a pool of more basic than tau protein in this region and, since the unpolymerized tubulin (Knops et al., 1991). In this interaction of these proteins with microtubules has been system, microtubule bundles are formed in 100% of the postulated to be one of an ionic nature (Vallee, 1982), infected cells using either three- or four-repeat tau- MAP2 would be more effective than tau protein. It has expressing virus. This result stresses the difficulties in been reported that MAP2 promotes assembly, nu- comparing tau protein's bundling activity where differ- cleates microtubules and stabilizes microtubules more ent expression systems have been used and indicates efficiently than tau protein (Sandoval and Vandekerck- that valid comparisons between different isoforms or hove, 1981). different fragments must be made within a single Recently, Lewis and Cowan (1990) have reported expression system. Moreover, it also shows that a that the area responsible for bundling microtubules in heightened level of protein expression can compensate MAP2 is the area homologous to amino acids 313-335 of for a lowered binding affinity towards inducing micro- four-repeat tau protein (282-304 in three-repeat) and tubule bundling. that MAP2 lacking this region is still able to bind Our proposal that tau induces microtubule bundling microtubules. In tau protein, this area is required not by stabilizing microtubules is also supported by our only for microtubule bundling but also for binding. This finding that the minimal unique tau sequence required suggests that the role of this area in bundling by MAP2 for binding and bundling is too small to span the 20-25 might be to enhance binding, perhaps by influencing the nm distance between bundled microtubules, especially conformation of the repeat region or by providing if the entire repeat region is bound to a single additional interactions with microtubules. In agreement microtubule. The possibility that a single repeat region with Lewis and Cowan (1990), we found no evidence binds to two microtubules would require that only a 236 G. Lee and S. L. Rook single repeat be in contact with each microtubule; given additional sequence on each side of the repeat region the much reduced binding affinity of single repeat units may be related to protein folding and/or to additional (Butner and Kirschner, 1991), it seems unlikely that microtubule contact sites. The presence of tau protein such a configuration would be physiologically relevant. in the cell is correlated with the acetylation of all The report that taxol produces similar microtubule cellular microtubules, possibly in a cell cycle-dependent bundles (Chapin et al., 1991) is consistent with our manner. When tau protein with four-repeats was findings. With the increased stability of microtubules in expressed, thick dense bands of microtubule bundles taxol-treated or tau-transfected cells, other cellular were observed in cells. Deletion analysis did not reveal factors may participate in the bundling process. Vallee any unique sequence associated with bundling; four- (1990) has raised the possibility that cells transfected repeat constructs capable of binding also induced with MAP2 may over-express other MAPs that may be microtubule bundling. Also, microtubules associating involved in bundling. Another possibility is that with four-repeat tau in vivo were found to have greater stabilized microtubules may have new physicochemical stability, as assessed by nocodazole resistance, than properties at their surfaces that make them self- those associating with three-repeat tau. Our data adherent. suggest that microtubule bundling is correlated with If microtubule bundling does not require other increased stability of microtubules. Since the fourth proteins, it should be reproducible in vitro. While we repeat is found only in adult isoforms of tau protein, our have been able to obtain microtubule bundles in vitro work provides evidence that the developmental regu- using E. co/i-synthesized tau protein in a centrosome- lation of tau protein heterogeneity bears functional mediated microtubule regrowth assay (Brandt and Lee, implications for cellular microtubules. unpublished data), the relationship between these bundles and those seen in 3T3 cells remains to be We sincerely thank Dr. Frank McKeon for help at various shown. stages of this work and Drs. Alfredo Caceres, Kenneth Kosik and David Gard for useful suggestions and discussion. We The presence of tau protein is correlated with the also thank Dr. Gianni Piperno for antibodies and Dr. Hiroshi acetylation of all cellular microtubules Mori for his cDNA clone. The contributions of Ms. Belinda The acetylation of microtubules has been postulated to Chang towards DNA sequencing are also gratefully acknowl- edged. Lastly, we thank Drs. David Gard, Garth Hall and be the consequence of microtubule stabilization Roland Brandt for critical reading of this manuscript. This (Piperno et al. 1987; Schulze et al., 1987; Webster and work was supported by the National Institute of General Borisy, 1989). Therefore, one might expect micro- Medical Sciences (no. GM39300-O4). tubules stabilized by tau protein to be acetylated. In our transfections, we found many transfected cells exhibiting an acetylated tubulin staining pattern coin- References ciding with the general cytoplasmic interphase microtubule array as detected by tau staining. This supplies Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, new evidence that tau protein stabilizes microtubules in J. G., Smith, J. A. and Struhl, K. (1989). Current Protocols in vivo, leading to the acetylation of all microtubules in Molecular Biology. John Wiley and Sons, New York, NY. Binder, L. I., Frankfurter, A. and Rebhun, L. I. (1985). The the cell; other lines of evidence have been based on the distribution of tau in the mammalian central nervous system. /. Cell use of microtubule depolymerizing drugs (Drubin and Biol. 101, 1371-1378. Kirschner, 1986). However, transfected cells with Brion. J. P., Guilleminot, J., Couchic, D., Flament-Durand, J. and acetylated tubulin staining patterns resembling those Nunez, J. (1988). Both adult and juvenile tau microtubule- associated proteins are axon specific in the developing and adult rat found in non-transfected cells were also present; in cerebellum. Neuroscience 25, 139-146. these cells, only a subset of microtubules were acety- Butner, K. A. and Kirschner, M. W. (1991). Tau protein binds to lated. This suggests that microtubule acetylation is microtubules through a flexible array of distributed weak sites. J. regulated by other factors in addition to stability, as Cell Biol. 115, 717-730. conferred by tau protein association. Webster and Caceres, A. and Kosik, K. S. (1990). Inhibition of neurite polarity by tau antisense oligonucleotides in primary cerebellar neurons. Borisy (1989) have found that acetylated microtubules Nature 341, 461-463. are absent among prophase microtubules and that their Caceres, A., Potrebfc, S. and Kosik, K. S. (1991). The effect of tau reappearance lags behind that of the cytoplasmic antisense oligonucleotides on neurite formation of cultured microtubules after the completion of mitosis. This cerebellar macroneurons. J. Neurosci. 11, 1515-1523. Chapin, S. J., Bulinski, J. C. and Gundersen, G. G. (1991). suggests that acetylation may be cell cycle regulated and Microtubule bundling in cells. Nature 349, 24. that our transfected cells have a distribution of Chen, C. C. and Okayama, H. (1988). Calcium phosphate-mediated acetylated tubulin patterns due to asynchrony in the gene transfer: a highly efficient transfection system for stably culture. transforming cells with plasmid DNA. Biotechniques 6, 632-638. Cleveland, D. W., Hwo, S.-Y. and Kirschner, M. W. (1977). Physical and chemical properties of purified tau factor and the role of tau in Conclusions microtubule assembly. J. Mol. Biol. 116, 227-247. In summary, our results show that regions outside the Drubin, D. G., Feinstein, S. C, Shooter, E. M. and Kirschner, M. W. repeat region of tau protein influence the ability of the (1985). Nerve growth factor-induced neurite outgrowth in PC12 cells involves the coordinate induction of microtubule assembly and protein to bind to microtubules in vivo. We define assembly-promoting factors. /. Cell Biol. 101, 1799-1807. regions flanking the repeats that are required for in vivo Drubin, D. G. and Kirschner, M. W. (1986). Tau protein function in binding under our conditions. The requirement for living cells. J. Cell Biol. 103, 2739-2746. Expression of tau protein in 3T3 cells 237

Drubin, D., Kobayashi, S., Kellogg, D. and Kirschner, M. (1988). short hydrophobic zipper in microtubule-associated proteins MAP2 Regulation of microtubule protein levels during cellular and tau. Nature 342, 498-505. morphogenesis in nerve growth factor-treated PC12 cells. J. Cell Lewis, S. A., Wang, D. and Cowan, N. J. (1988). Microtubule- Biol. 106, 1583-1591. associated protein MAP2 shares a microtubule binding motif with Ellis, L., Clauser, E., Morgan, E. O., Edery, M., Roth, R. A. and tau protein. Science 242, 936-939. Rutter, W. J. (1986). Replacement of insulin receptor tyrosine LindwaU, G. and Cole, R. D. (1984). Phosphorylation affects the residues 1162-1163 comprises insulin-stimulated kinase activity and ability of tau protein to promote microtubule assembly. /. Biol. uptake of 2-deoxyglucose. Cell 45, 721-732. Chem. 259, 5301-5305. Ennulat, D. J., Llem, R. K. H., Hashim, G. A. and Shelanski, M. L. Mori, H., Hamada, Y., Kawaguchi, M., Honda, T., Kondo, J. and (1989). Two separate 18-amino acid domains of tau promote the Ihara, Y. (1989). A distinct form of tau is selectively incorporated polymerization of tubulin. J. Biol. Client. 264, 5327-5330. into Alzheimer's paired helical filaments. Biochem. Biophys. Res. Goedert, M. and Jakes, R. (1990). Expresson of separate isoforms of Commun. 159, 1221-1226. human tau protein: correlation with the tau pattern in brain and O'Farrell, P. Z., Goodman, H. M. and O'Farrell, P. H. (1977). High effects on tubulin polymerization. EMBO J. 9, 4225-4230. resolution two-dimensional electrophoresis of basic as well as acidic Goedert, M., Spillantini, M. G., Jakes, R., Rutherford, D. and Crowther, R. A. (1989b). Multiple isoforms of human microtubule- proteins. Cell 12, 1133-1142. associated protein tau: sequences and localization in neurofibrillary Olmsted, J. B. (1986). Microtubule-associated proteins. Annu. Rev. tangles in Alzheimer's disease. Neuron 3, 519-526. Cell Biol. 2, 421-457. Goedert, M., Spillantini, M. G., Potier, M. C, Ulrich, J. and PfefTer, S. R., Drubin, D. G. and Kelley, R. B. (1983). Identification Crowther, R. A. (1989a). Cloning and sequencing of the cDNA of three coated vesicle components as a- and /J-tubulin linked to a encoding an isoform of microtubule-associated protein tau phosphorylated 50,000-dalton polypeptide. J. Cell Biol. 97, 40-47. containing four tandem repeats: differential expression of tau Piperno, G. and Fuller, M. T. (1985). Monoclonal antibodies specific protein mRNAs in human brain. EMBO J. 8 393-399. for an acetylated form of a^tubulin recognize the antigen in cilia Goedert, M., Wischik, C. M., Crowther, R. A., Walker, J. E. and and flagella from a variety of organisms. J. Cell Biol. 101, 2085- Klug, A. (1988). Cloning and sequencing of the cDNA encoding a 2094. core protein of the paired helical filament of Alzheimer disease: Piperno, G., LeDizet, M. and Chang, X. (1987). Microtubules Identification as the microtubule-associated protein tau. Proc. Nat. containing acetylated o^tubulin in mammalian cells in culture. J. Acad. Sci. USA 85, 4051-4055. Cell Biol. 104, 289-302. Himmler, A. (1989). Structure of the bovine tau gene: alternatively Sandoval, I. V. and Vandekerckhove, J. S. (1981). A comparative spliced transcripts generate a protein family. Mol. Cell. Biol. 9, study of the in vitro polymerization of tubulin in the presence of the 1389-1396. microtubule-associated proteins MAP2 and x. J. Biol. Chem. 256, Himmler, A., Dreschel, D., Kirschner, M. W. and Martin, D. W., Jr 8795-8800. (1989). Tau consists of a set of proteins with repeated C-terminal Schulze, E., Asal, D. J., Bulinski, J. C. and Kirschner, M. W. (1987). microtubule-binding domains and variable N-terminal domains. Posttranslational modification and microtubule stability. J. Cell Mol. Cell. Biol. 9, 1381-1388. Biol. 105, 2167-2177. Johnson, G. D. and Araujo, G. M. de C. N. (1981). A simple method Solomon, F., Magendantz, M. and Sal/man, A. (1979). Identification of reducing the fading of immunofluorescence during microscopy. with cellular microtubules of one of the co-assembling microtubule- J. Immunol. Meth. 43, 349-350. associated proteins. Cell 18, 431-438. Joly, J. C, Flynn, G. and Purlch, D. L. (1989). microtubule binding Steiner, B., Mandelkoow, E.-M., Biernat, J., Gustke, N., Meyer, H. fragment of microtubule-associated protein 2: location of the E., Schmidt, B., Mieskes, G., Soling, H. D., Drechsel, D., protease- accessible site and identification of an assembly Kirschner, M. W., Goedert, M. and Mandelkow, E. (1990). promoting peptide. J. Cell Biol. 109, 2289-2294. Phosphorylation of microtubule-associated protein tau: Kanai, Y., Takemura, R., Oshima, T., Mori, H., Ihara, Y., identification of the site for calcium-calmodulin dependent kinase Yanagisawa, M., Masaki, T. and Hlrokawa, N. (1989). Expression and relationship with tau phosphorylation in Alzheimer tangles. of multiple tau isoforms and microtubule bundle formation in fibroblasts transfected with a single tau cDNA. J. Cell Biol. 109, EMBO J. 9, 3539-3544. 1173-1184. Studler, F. W., Rosenberg, A. H., Dunn, J. J. and DubendorfT, J. W. Knops, J., Koslk, K. S., Lee, G., Pardee, J. D., Cohen-Gould, L. and (1990). Use of T7 RNA polymerase to direct the expression of McConlogue, L. (1991). Overexpression of tau in a nonneuronal cloned genes. Meth. Enzymol. 185, 60-89. cell induces long cellular processes. J. Cell Biol. 114, 725-733. Trojanowski, J. Q., Schuck, T., Schmidt, L. and Lee, V. M.-Y. Kowall, N. W. and Koslk, K. S. (1987). Axonal disruption and (1989). Distribution of tau proteins in the normal human central aberrant localization of tau protein characterize the neuropil and peripheral nervous system. J. Histochem. Cytochem. 37, 209- pathology of Alzheimer's disease. Ann. Neurol. 22, 639-643. 215. Laemmli, U. K. (1970). Cleavage of structural proteins during the Vallee, R. B. (1982). A taxol-dependent procedure for the isolation of assembly of the head of bacteriophage T4. Nature 227, 680-685. microtubules and microtubule-associated proteins (MAPs). J. Cell Lee, G., Cowan, N. and Kirschner, M. (1988). The primary structure Biol. 92, 435-442. and heterogeneity of tau protein from mouse brain. Science 239, Vallee, R. B. (1990). Molecular characterization of high molecular 285-288. weight microtubule-associated proteins: some answers, many Lee, G., Neve, R. L. and Kosik, K. S. (1989). The microtubule questions. Cell Motil. Cytoskel. 15, 204-209. binding domain of tau protein. Neuron 2, 1615-1624. Webster, D. R. and Borisy, G. G. (1989). Microtubules are acetylated Lewis, S. A. and Cowan, N. (1990). Microtubule bundling. Nature in domains that turn over slowly. J. Cell Sci. 92, 57-65. 345, 674. Lewis, S. A., Ivanov, 1. E., Lee, G.-H. and Cowan, N. J. (1989). Microtubule organization in dendrites and axons is determined by a {Received 23 January 1992 - Accepted 9 March 1992)