Cell Motility and the Cytoskeleton 21:31%325 (1992)
Differential Sorting of Beta Tubulin lsotypes Into Colchicine-Stable Microtubules During Neuronal and Muscle Differentiation of Embryonal Carcinoma Cells Marcia M. Falconer, Christophe J. Echeverri, and David L. Brown Department of Biology, University of Ottawa, Ottawa, Canada Pluripotent PI 9 embryonal carcinoma (EC) cells were differentiated along the neuronal and muscle pathways. Comparisons of class I, 11,111, and IV beta tubulin isotypes in total and colchicine-stable microtubule (MT) arrays from uncommitted EC, neuronal, and muscle cells were made by immunoblotting and by indirect immunofluorescence microscopy. In undifferentiated EC cells the relative amounts of these four isotypes are the same in both the total and stable MT populations. Subcellular sorting of beta tubulin isotypes was demonstrated in both neuronal and muscle differentiated cells. During neuronal differentiation, class I1 beta tubulin is preferentially incorporated into the colchicine-stable MTs while class I11 beta tubulin is preferentially found in the colchicine-labile MTs. The subcellular sorting of class I1 into stable MTs correlates with the increased staining of MAP l B , and with the expression of MAP 2C and tau. Although muscle differentiated cells express class I1 beta tubulin, stable MTs in these cells do not preferentially incorporate this isotype but instead show increased incorporation of class IV beta tubulin. Muscle cells do not show high levels of MAP 1B and do not express MAP 2C or tau. These results are consistent with the hypothesis that a subcellular sorting of tubulin isotypes is the result of a complex interaction between tubulin isotypes and MT-associated proteins. Key words: pluripotent P19 EC cells, immunoblotting, indirect immunofluorescencemicroscopy, microtubule-associated proteins, MAP2, tau, MAP 1B
INTRODUCTION
Microtubules (MTs) are structurally similar at the level of the electron microscope, but are involved in many distinct functions including maintenance of cellular architecture, intracellular transport, cell division, and cell motility. What determines which MTs perform a particular function remains unanswered. The discovery of multiple isoforms for alpha and beta tubulin, which together form the tubulin heterodimer, suggested that tubulin variants, perhaps in combination with specific MT-associated proteins (MAPS) as suggested by Kim et al. [1986], Lewis and Cowan [1990b], and others, may be involved in MT selectivity [Fulton and Simpson, 19761. However, evidence to support this hypothesis has been unexpectedly difficult to achieve. In higher vertebrates, sequencing of tubulin genes has shown the presence of about 6 alpha and 6 beta 0 1992 Wiley-Liss, Inc.
tubulin isotypes [reviewed by Joshi and Cleveland, 1990; Lewis and Cowan, 1990b; Sullivan, 19881. In beta tubulins, and to a lesser extent in alpha tubulins, variable regions at the carboxy terminus of each isotype are highly conserved across species lines suggesting that these isotypes have specific functions. The final 15 amino acids of the carboxy variable regions have been used to classify beta tubulin isotypes [Lopata and Cleveland, 1987; reviewed by Sullivan, 19881 and both poly-
Received July 24, 1991; accepted October 28, 1991 Address reprint requests to David L. Brown, Dept. of Biology, Univ. of Ottawa, Ottawa, Canada K1N 6N5.
M.M. Falconer is now at Dept. of Biology, M.I.T., E17-221, Cambridge, MA.
314
Falconer et al.
clonal and monoclonal antibodies have been prepared and used to localize MTs containing these isotypes [Lopata and Cleveland, 1987; Lewis et al., 1987; Banerjee et al., 1988; Lee et al., 1990bl. In the relatively unspecialized cells of tissue culture cell lines, indirect immunofluorescence observations showed that all MTs are polymers of all available isotypes [Lopata and Cleveland, 1987; reviewed by Lewis and Cowan, 199Obl. However, examination of PC 12 neurons and of rat and squid axons has shown preferential isotype usage within the cell [Arai and Matsumoto, 1988; Joshi and Cleveland, 1989; Asai and Remolona, 1989; Denoulet et al., 19891. Neuronal differentiation is accompanied by an increase in MT stability [Black and Greene, 1982; Black et al., 1986; Lim et al., 19891 and by changes in both alpha and beta tubulin isoforms. Increased detyrosination and acetylation of alpha tubulin have been reported in stable neuronal MTs [Cambray-Deakin and Burgoyne, 1987; Black and Keyser, 1987; Ferreira and Caceres, 1989al. During neuronal differentiation in the P 19 embryonal carcinoma cell line, these post-translational modifications occur within the first 24 hours of differentiation and are the earliest evidence of increased MT stability [Falconer et al., 1989a,b]. However, acetylation and detyrosination of alpha tubulin do not appear to be the cause of increased MT stability [Schulze et al., 1987; Khawaja et al., 1988; Webster et al., 19901, indicating that beta tubulin isotypes may be important in this role. Changes in beta tubulin isotype expression have been shown in differentiating PC 12 neurons and regenerating nerves [Joshi and Cleveland, 1989; Hoffman and Cleveland, 19881. To examine the suggestion that beta tubulin isotypes are important to MT stability we use the multipotent, murine PI9 embryonal carcinoma (EC) cell line which can be induced to differentiate along the neuronal or muscle pathway by varying the morphogen [JonesVilleneuve et al., 1982; Edwards et al., 1983; McBurney et al., 1988; Rudnicki et al., 19901. The expression and changes in the relative percentages of beta tubulin isotypes are monitored using indirect immunofluorescence microscopy and immunoblotting, in both the total MT array and in the colchicine-stable MT array of undifferentiated EC cells, and cells differentiating into neurons or muscles. Only beta tubulin isotype classes I, 11, 111, and IV are included in this study; mouse cells do not express class V and the P19 EC cell line did not express class V as judged by immunoblotting of whole cell extracts and by indirect immunofluorescence microscopy using a polyclonal antibody to a peptide representing the COOH terminus of class V beta tubulin [class VI is specific to haematopoetic cells, Sullivan, 19881. Our results show that colchicine-stable MTs in un-
differentiated EC cells exhibit no preferential isotype usage compared with the isotypes in the total MT population; however, during neuronal differentiation, subcellular sorting of two beta tubulin isotypes occurs. Class I1 beta tubulin is preferentially included in neuronal stable MTs while class 111 is preferentially included in the colchicine-labile MT population. The subcellular sorting of class I1 beta tubulin coincides with, and may be related to, the sequential expression of MAP lB, MAP 2, and tau. The beta tubulin isotype profile of EC cells differentiated along the muscle pathway is distinct from that of both neuronal and undifferentiated EC cells. Colchicinestable MTs in muscle cells do not show increased sorting of class I1 beta tubulin, although there is sorting of class IV beta tubulin into stable MTs compared to the total MT array. In addition, cells differentiated along the muscle path do not express brain-specific MAP 2 and tau. MATERIALS AND METHODS Cell Culture
Stock cultures of the P19 murine embryonal carcinoma cell line [McBurney and Rogers, 19821 were maintained at 37°C and 5% CO, in alpha-MEM (Flow Lab., Mississauga, Ont.), supplemented with 10%heat-inactivated fetal bovine serum (FBS) (Flow), and passaged every 2 days. Neuronal Differentiation
For neuronal differentiation, stock cultures were trypsinized, resuspended in alpha-MEM plus 10% FBS at the appropriate concentration (see below), aggregated in a microbiology grade Petri dish for 1 h at 37°C and 5% CO,, then plated in 100 mm Corning tissue culture dishes. After 24 h, medium was replaced with alphaMEM plus 2% FBS containing lop6 M retinoic acid (RA) (Sigma, St. Louis, MO) and the cultures were maintained in this medium for the total differentiation period. Differentiating cells were observed in replicate plates after 1, 2, 4, and 6 days of RA treatment. Plating densities varied from 1 X lo4 cellddish for cells differentiated for 4 or 6 days to 1 X lo6 cells/dish for cells differentiated for 1 or 2 days. RA was prepared as a lo-' M stock solution dissolved in ethanol, stored at -8O"C, and appropriate amounts were added to the medium for a final concentration of M RA. Muscle Cell Differentiation
For differentiation of muscle cells, aggregated cells were plated at 1 X lo4 cells/100 mm tissue culture dish in alpha-MEM plus 10% FBS. Approximately 12 h after plating, the medium was replaced with alpha-MEM plus 8% FBS and 2% dimethylsulfoxide (DMSO) (Sigma).
Differential Beta Tubulin Isotype Sorting
Cells were maintained in this medium for 6 days of differentiation. Cell Extracts for lmmunoblotting
To prepare cell extracts containing tubulin from the total polymerized MT population, a 100 mm tissue culture dish containing a semi-confluent monolayer of either untreated EC cells, or cells treated for 2 , 4 , or 6 days with RA, or cells treated with DMSO for 6 days was washed 1 X with 37°C PBS and 2 X with 37°C MT stabilizing PEM buffer (80 mM PIPES, 5 mM EGTA, 1 mM MgC1, [all from Sigma] pH 6.8). Cells were then permeabilized for 3 to 5 rnin in PEM buffer plus 0.75% Triton X- 100 (Sigma) at 37°C to extract non-polymerized tubulin and other soluble proteins. The remaining cytoskeleton, containing the total MT array, was solubilized in 1 ml of 0.5% SDS (Bio-Rad, Rockville Centre, NY) in 25 mM Tris (Bio-Rad) pH 6.8 and 2 M glycerol, by boiling for 10 min in a microfuge tube. The extracts were centrifuged at 13,000 rpm for 3 min, and stored at -20°C. Prior to use, the sample was thawed, centrifuged at 13,000 rpm for 3 min, and total protein in the supernatant was determined (Bio-Rad protein determination kit catalog number 500-0002). For polyacrylamide gel electrophoresis and immunoblotting, equal amounts of sample proteins were diluted in gel sample buffer and loaded, usually 100 pg/well. Samples containing tubulin from the colchicine-stable MT population were obtained as described above except that the cells were treated with 1 pg/ml colchicine (Sigma) for 45 min before extraction. Three or more independent samples were examined for each day of differentiation with the exception of 2 days RA neuronal differentiation in which only two independent samples were examined.
315
?-labeled goat anti-rabbit (specific activity 7.7 1 pCi/ pg) or rabbit anti-mouse (specific activity 8.8 1 pCi/pg) secondary antibodies (New England Nuclear, DuPont Canada, Mississauga, Ont.). ‘251-labeled secondary antibodies were used at 5 X lo5 c p d m l and incubated overnight followed by 4 X 20 rnin washes in PBS plus 0.1% Tween-20. Secondary controls in which the primary antibody was omitted were included in each experiment. Binding was detected using Kodak X-OMAT film exposed at -80°C for 3 days before development. X-ray films were scanned using an LKB Ultroscan XL laser densitometer which computed the area under the curve. Each band was scanned a minimum of 5 times at 800 pm intervals. Scans produced a bell shaped curve reflecting lower levels of radioactivity at the edges of each band. The increasing/decreasing values on each side of the bell curve were omitted and only the levels reflecting the peak of the curve were averaged and used. A tubulin dilution series was immunoblotted with the 5A6 antibody which recognizes most forms of alpha tubulin [Aitchison and Brown, 19861 and scanned to determine that the experimental densitometer readings were in the linear range. The densitometer values of all isotypes in any one sample are expressed as relative percentages of the total isotypes measured and should not be interpreted as actual amounts of isotype in the sample. To calculate the relative percentage of isotype x, where “area x ’ ’ is the computed area under the curve found by laser densitometric scanning: (area X x 100%)/(area isotype I) + (area isotype 11) (area isotype 111) (area isotype IV) = relative percentage of isotype X. Student’s t test was applied to determine if sample means differed significantly at the 95% confidence level.
+
+
Gel Electrophoresis and lmmunoblotting
Beta tubulin isotypes in the total polymerized MT array and in the colchicine-stable MT array of undifferentiated EC cells, and neuronal and muscle differentiated cells were examined using immunoblotting . The samples were loaded as “curtain gels” on 7.5% polyacrylamide (Bio-Rad) gels and run at 175 volts for 45 rnin using a Bio-Rad Mini-gel system. Proteins were electrophoretically transferred overnight at 30 milliamps to nitrocellulose membrane (BA 85; Schleicher & Schuell, Inc.) in Laemmli gel running buffer containing 20% methanol. Nitrocellulose filters were stained for 5 rnin with Ponceau red (0.2% Ponceau S dye [Sigma] w/v in 3% trichloroacetic acid [Sigma]) then destained in PBS. Filters were blocked in 5% skim milk powder (Carnation) dissolved in PBS for 1 h. For immunoblotting, a Miniblotter 28 (Immunetics, Cambridge, MA) was used. Primary antibodies, diluted in PBS, were incubated overnight followed by 4 X 30 rnin washing in PBS before addition of
Antibodies Used
The primary antibodies used for immunoblotting and/or indirect immunof luorescence staining were generously provided by the following: rabbit polyclonal antibodies specific to beta tubulin isotype classes I, 11, IV, and V from D.W. Cleveland [Lopata and Cleveland, 19871; mouse mAb to class I11 beta tubulin, TuJ1, from A. Frankfurter [Lee et al., 1990al; mouse mAb to class I1 beta tubulin, from R.L. Luduena [Banerjee et al., 19881; rabbit polyclonal antibody to detyrosinated alpha tubulin from Dr. J.C. Bulinksi and Dr. G.G. Gundersen [Gundersen et al., 19841; Mouse mAbs to MAP 1B (1W6D4), tau (Tau-1), and MAP 2 (clones AP-14 and AP18) [Tucker and Matus, 19881 from L.I. Binder. Rat monoclonal YOL 1/34 (Dimension Lab. Mississauga, Ont.) was used as a general anti-tubulin stain [Kilmartin et al., 19821. Mouse mAb to muscle-specific a and y actins, HHF35 (MA-931) [Tsukada et al.,
316
Falconer et al.
19871, was used to identify muscle-differentiated cells (Enzo Biochem. Inc. New York). Indirect ImmunofIuorescence Microscopy
Coverslips coated with 0.1% sterile gelatin (Sigma), were placed in the 100 mm tissue culture dishes before addition of cell aggregates. For observations of colchicine-stable MTs, cells were treated with 1 pg/ml colchicine for 45 rnin before fixation. For indirect immunofluorescence staining of MTs, cells grown on coverslips were washed with PBS then simultaneously fixed and extracted in 3.7% paraformaldehyde plus 0.25% glutaraldehyde (v/v) in PEM buffer plus 1% Triton X-100 for 10 min then rinsed 3 X 3 min in PBS. For MAP staining, cells were fixed in cold MeOH at -20°C for 10 to 15 rnin then rinsed 3 X 5 rnin in PBS. Fixed cells were incubated with primary antibody for 45 min, rinsed 3 X 5 min in PBS, incubated 45 rnin in secondary antibody, stained for 1 rnin in 1 pg/ml Hoechst dye #33258 to visualize nuclei, rinsed 3 x 5 rnin in PBS and mounted for observation. Double immunofluorescence staining was done by simultaneous incubation in two primary antibodies followed by simultaneous incubation in two secondary antibodies. Appropriate controls were done to show that neither crossreaction among secondary antibodies nor bleed-through occurred. Secondary antibodies for microscopy were FITCconjugated goat anti-rat IgG cross-absorbed against mouse, FITC-conjugated goat anti-mouse IgG cross-absorbed against rat, FITC-conjugated donkey anti-rabbit IgG cross-absorbed against rat and mouse, rhodamineconjugated rabbit anti-mouse IgG cross-absorbed against rat all from Chemicon, Temecula, CA; and rhodamineconjugated goat anti-mouse IgG cross-absorbed against rat (Jackson Inc. West Grove, PA). RESULTS
The P19 EC cell line is pluripotent and differentiation can be directed along the neuronal pathway by addition of RA or along the muscle pathway by addition of DMSO allowing us to examine and compare changes in the beta tubulin isotypes of the total and colchicinestable MT populations in three cell types: undifferentiated EC cells, differentiating neuronal cells, and differentiating muscle cells. Changes in classes I, 11, 111, and IV beta tubulin are examined, concentrating on class 11, a major brain isotype which is also present in other tissues, and class 111, which is neuron specific in brain [Sullivan, 1988; Lee et al., 1990al. Immunoblotting results are expressed as relative percentages of classes I, 11, 111, and IV beta tubulin isotypes within each sample. This allows comparison of
isotype expression in differentiating cells with that present in the undifferentiated EC cells, but does not reflect actual amounts of the isotypes. To simplify discussion, histograms of the beta tubulin isotypes in a given sample will be defined as the “isotype profile” for that sample. We examined isotype profiles for both the total MT array and for the colchicine-stable MT array in each of the three differentiation states. Undifferentiated EC Cells
In undifferentiated EC cells, the isotype profiles of the total MT array (Fig. la) and of the colchicine-stable MT array (Fig. lb) are essentially identical. Indirect immunofluorescence staining of EC cells, by antibody to class TI or by the TuJl antibody to class I11 beta tubulin, shows faint staining of MTs in most cells (not shown). Neuronal Differentiation
1 day RA. Immunostaining shows increased expression of class I1 isotype in a subset of cells after 1 day of differentiation, and, at the level of the light microscope, there appears to be preferential incorporation of this isotype into MTs which form a bundle (Fig. 2a,b). In colchicine-treated cells, class I1 beta tubulin is present in bundled stable MT arrays which also are found only in a subset of cells (Fig. 2c,d). Expression of class 111 isotype apparently is reduced below the level in EC cells, as indicated by loss of staining with TuJl antibody (not shown), and stable MTs in colchicine-treated cells show no staining with TuJl antibody to class I11 beta tubulin (Fig. 2e,f). 2 days RA. Cells differentiated for 2 days show a significant increase in the relative percentage of class I1 as well as a decrease in class IV in both the total and stable MT arrays (see Fig. la,b). Immunostaining indicates that all cells have a higher level of class I1 expression after 2 days of differentiation compared to the staining level after 1 day differentiation; however, a subset of cells, positioned on top of the monolayer, stains more intensely than do the cells of the monolayer. These cells with higher class I1 levels often have short processes (Fig. 3a,b). MTs in these cells and in colchicine-treated cells stain crisply with antibody to class I1 beta tubulin (Fig. 3b). There also is increased expression of class 111 beta tubulin in a subset of cells on top of the monolayer. Staining with the TuJl antibody to class I11 beta tubulin shows both MT staining as well as diffuse, intracellular fluorescence (Fig. 3c). In colchicine-treated cells, class 111 beta tubulin appears almost entirely as a brilliant, diffuse intracellular stain with few or no discernable MTs (Fig. 3d). 4 to 6 days RA. After 4 days differentiation, the isotype profiles of both the total and the stable MT populations show a large increase in the relative percentage
Differential Beta Tubulin Isotype Sorting
a, 1class I
1class 1class 1 II
111
ClasSIV
70 60
50 40
30 20 10
C EC
b. 0class I
D2
1
class I I
[1class 111
class IV
76001 50
r
40
317
of class 111, neuron-specific beta tubulin. However, a comparison of total and stable tubulin profiles shows that the stable MTs are enriched for class I1 and have a lower relative percentage of class 111 beta tubulin than does the total MT array (Fig. la,b). Although the amount of neuronal differentiation varies from experiment to experiment, by 6 days of differentiation, up to 80% of the cell population has long, neurite-like extensions. The remainder of the population consists of cells which stain with the MA-931 antibody to muscle-specific actin, fibroblast-like cells, and cells with EC-like morphology. At this point, the beta tubulin profiles of the total and the colchicine-stable MT populations are significantly different (compare Fig. la,b). Stable MTs are enriched in class I1 beta tubulin while the total MT population, which largely represents the colchicine-labile MTs, is enriched in class 111. During the period between 2 and 6 days of neuronal differentiation, immunostaining shows that cells on top of the cell monolayer extend neurite-like processes. The expression of class I1 beta tubulin increases in this subset of cells and decreases in the cells of the underlying monolayer (Fig. 4a). Stable MTs in colchicine-treated cultures stain crisply with antibody to class 11, mainly visible in cell bodies and neurites (Fig. 4c). Unlike class I1 beta tubulin, expression of class I11 is limited to a subset of differentiating neuronal cells; cells of the monolayer do not stain with the TuJl antibody (Fig. 4b). The differentiating cells stained with TuJl continue to show a high level of diffuse intracellular fluorescence, especially in colchicine-treated cells. Colchicine-treated cells now show class I11 beta tubulin staining of some stable MTs in the cell body; however, any stable MTs in the neurites are obscured by high levels of intracellular fluorescence (Fig. 4d). MAPs
30
20
10 0
EC
D6
Fig. 1 . Isotype profiles showing sequential changes in the relative percentages of beta tubulin isotypes in (a) the total MT array and (b) the stable MT array of neurally differentiating P19 EC cells. Mean and standard deviation based on the following number of independent samples: for total MTs-(EC) 5 samples, (2 day RA) 2 samples, (4 day RA) 3 samples, (6 day RA) 4 samples; for stable MTs--(EC) 4 samples, (2 day RA) 2 samples, (4 day RA) 3 samples, (6 day RA) 3 samples.
We used antibodies to MAP l B , MAP 2, and tau to examine the appearance of these MAPs during differentiation. Low level MAP 1B staining (Fig. 5a), but not MAP 2 or tau can be detected in all uncommitted EC cells fixed directly in MeOH at -20°C. After 2 days differentiation, MAP 1B increases in a subset of cells, in a pattern similar to that of class I1 beta tubulin (compare Fig. 5b with Fig. 2a) and by 6 days differentiation MAP 1B is present in neurite-like processes (Fig. 5c). Previously we reported that differentiating EC cells did not stain for MAP 1B [Falconer et al., 1989al. This was an artifact of simultaneous fixatiodextraction with detergent [see Connolly and Kalnins, 19801. The earliest brain-specific MAP detected during neuronal differentiation is the 70 kD form of MAP 2, MAP 2C. Identification was by comparison of staining patterns of two antibodies, AP- 14, which recognizes
318
Falconer et al.
Fig. 2. Neural differentiation after 1 day in RA. a: Mouse mAb to class I1 beta tubulin shows increased staining in a subset of cells. Compare with b, a double label of same cells as in a using rat mAb YOL 1/34 as a general MT stain; c: Stable MTs in colchicine-treated cells stained with mouse mAb to class 11. Compare with d, double label of same cells as in c using rabbit polyclonal antibody to dety-
rosinated alpha tubulin to identify stable MTs. e: No staining with mAb TuJl to class I11 beta tubulin is detected in stable MTs of colchicine-treated cells. Compare with f, double label of same cells as in e using rabbit polyclonal antibody to detyrosinated alpha tubulin to identify stable MTs. Bar, 20 pm.
only the high molecular weight form of MAP 2, and AP- 18, which recognizes both low and high molecular weight MAP2 forms [see Tucker and Matus: 19881. During the first 6 days of neuronal differentiation in EC cells, MAP 2 immunostaining can be detected by AP-18, but not by AP-14. Immunoblotting shows that low levels of MAP 2C can be detected beginning at 2 days differentiation (2 days RA) and that high molecular weight MAP 2 can be detected beginning at 6 days RA (Fig. 6a). We cannot exclude the possibility that high molecular weight MAP 2 is expressed earlier but at levels below the limit of detection for the methods used. Staining with AP-14 is found after 7 or more days of differentiation. Immunostaining indicates that tau protein, as detected by the Tau-1 antibody, is first seen between days 4 and 6 of differentiation. This is confirmed by immunoblots which show the presence, at 6 days of differentiation, of a single band with an apparent molecular weight of about 48 kD, indicative of a juvenile form of tau (Fig. 6b). When EC cells are differentiated for 6 days along the muscle pathway, MAP 2, 2C, and tau are not detected using either immunostaining or immunoblotting.
Only low level staining of MAP 1B is found in muscle differentiated cultures.
Beta lsotypes in
and Differentiating The sorting of class I1 beta tubulin into neuronal stable MTs coincides with the appearance of MAP 2C. Isotype sorting increases coordinately with expression of MAP 2C and tau. In addition, undifferentiated EC cells do not express either of these MAPs and do not show isotype differences between the total and stable MT arrays. This data indicated to us that subcellular sorting may depend upon a complex interaction between brainspecific MAPs and beta tubulin isotypes. Therefore we examined the isotype profiles and brain MAPs in muscle differentiating cells and compared these to the isotype profiles in neuronal and EC cells. The amount of muscle differentiation varies between experiments, but in the best case up to 60% of cells differentiated for 6 days in DMSO stain with an antibody which recognizes a and y muscle-specific actins, HHF35 (Fig. 7a,b). The remaining cells have fibro-
Differential Beta Tubulin Isotype Sorting
319
Fig. 3. Neural cells differentiated for 2 days in RA and stained with mouse mAb to class I1 beta tubulin (a,b) and with TuJl mAb to class 111 beta tubulin (c,d). a: A subset of cells shows increased expression of class I1 beta tubulin with preferential incorporation into MTs in cell processes and/or neurites. Note that cells in the underlying monolayer also show class I1 staining but at a lower intensity than the differentiating subset of cells. b: In colchicine-treated cells, antibody to class
I1 isotype stains stable MTs in a subset of cells. c: There is induction of class I11 expression in a subset of cells which now stain positively with TuJl . Note that cells in the underlying monolayer do not express class 111. d: In colchicine-treated cells, antibody to class 111 isotype shows a strong but diffuse intracellular fluorescence which conceals any stable MTs. Bar, 20 km.
blast-like or EC-like morphology. No cells with neuritelike extensions are detected. When these cells are double labeled with YOL 1/34, as a general MT stain, and antibody to class I1 beta tubulin, all MTs contain a significant amount of class I1 beta tubulin (Fig. 7c,d). A comparison of beta tubulin isotype profiles in total and stable MT arrays of muscle, neuronal, and undifferentiated EC cells can be seen in Figure 8a,b. In neuronal cells, the total MT array contains only a low relative percentage of class I1 beta tubulin; nevertheless, neuronal stable MTs have a high relative percentage of this isotype. Muscle cells do not show this pattern. The profile of the total MT array in muscle cells contains a high relative percentage of class I1 beta tubulin, but the profile of the stable MTs does not show selective incorporation of this isotype. In addition, there is a significant increase in the percentage of class IV beta tubulin in the stable MT array (at the 90% confidence level).
MTs and formation of a colchicine-stable MT array [Falconer et al., 1989al. Neurons differentiated for 6 days have elongated neurites, express neurofilament proteins 160 and 68, synthesize the excitatory neurotransmitter ACh, and have high-affinity uptake sites for the inhibitory neurotransmitter GABA [Jones-Villeneuve et al., 1982; McBurney et al., 19881; however, differentiation is not complete and it is likely that there are further changes in tubulin isotype expression or sorting beyond those reported here. MT Stability, Beta Tubulin lsotypes and MAPS
The first observed increase in MT stability could arise by interaction of MAP 1B and class I1 beta tubulin, both of which show increased levels of expression in a subset of cells by day 2 of differentiation. This is similar to differentiatingPC12 cells which also show an increase in MAP 1B and in MT stability [Drubin et al., 19853 as well as a large increase in class I1 beta tubulin [Joshi and Cleveland, 19891. These patterns parallel the expression DISCUSSION of MAP 1B and class I1 beta tubulin in brain developTubulin lsoforms and MT Stability ment and nerve regeneration [Riederer et al., 1986; HoffDuring neuronal differentiation of P19 EC cells, man and Cleveland, 1988; Diaz-Nido and Avila, 1989; MTs become progressively resistant to depolymerizing Garner et al., 19901. agents [Wasteneys et al., 1988; Cadrin et al., 1988; FalWe suggest that the subsequent increase in MT staconer et al., 1989al and this increased stability is paral- bility and correlative subcellular sorting of class I1 beta leled by increasing acetylation and detyrosination of tubulin may be due to interaction with newly expressed
320
Falconer et al.
Fig. 4. Immunostaining of cells after 4 days of neuronal differentiation. (a and c stained with mouse mAb to class I1 beta tubulin, b and d stained with mAb TuJl to class 111.) a: Class I1 beta tubulin in MTs of the cell body and neurite-extensions stain with much greater intensity than do MTs in cells of the underlying monolayer. b: Differentiating neurons show clear class I11 beta tubulin-positive MTs as well as
diffuse, intracellular fluorescence. c: In colchicine-treated cultures, stable MTs containing class I1 are visible in neuronal cell bodies, and long neurites. d: A colchicine-treated neuron stained for class 111 shows diffuse but strong intracellular fluorescence and some stable MTs in the cell body. Bars in a, b: 20 pm. Bars in c, d: 10 pm.
MAP 2C and tau. MAP 2 is associated with neurite extension. Transfection of a MAP 2 anti-sense construct into the P19 EC cell line demonstrates that inhibition of MAP 2 early in neuronal differentiation greatly reduces the number of neurites which develop [Dinsmore and Solomon, 19911. Although a relationship between MT stability and the presence of both MAP 2 and tau has been demonstrated, nothing is known of the beta tubulin isotypes involved [Ferreira et al., 1989b; Kanai et al., 1989; Lewis et al., 1989; Lewis and Cowan, 1990al. In vitro evidence indicates that the amino acid sequence GEFEEEG, in positions 434-440,has the highest affinity for both MAP 2 and tau and that this sequence is found only in class I1 beta tubulin [Littauer et al., 1986; Luduena et al., 1988; Banerjee et al., 19901. The class 111 amino acid sequence in these positions is EMYEDDE and may have lower affinity for MAPs as suggested in Littauer et al. [1986], Luduena et al. [1988], and Banerjee et al. [1990]. MT-MAP interactions could result in formation of two distinct sets of MTs; a stable, structural array of MTs high in class I1 beta tubulin and rich in MAPs and, a dynamic, probing array of MTs high in class 111 beta tubulin and with fewer MAPs bound to them.
Class II and Class 111 Beta Tubulin
In vitro experiments indicate that preferential incorporation of a single isotype from a mixture of isotypes can occur [Murphy, 19881. If this occurs in vivo, MTs which acquire a high level of class I1 beta tubulin when gene expression is increased at 1 day differentiation may continue to preferentially incorporate more class TI. Moreover, if class I1 beta tubulin has a greater affinity for MAPs, then MTs containing higher levels of class I1 will bind more MAPs and become more stable, which allows incorporation of more class I1 in a self-perpetuating cycle. Two posttranslational modifications of class 111 beta tubulin, phosphorylation and polyglutamylation, may also play a role in isotype sorting. Phosphorylation occurs only on class 111 beta tubulin at serine 444 [Gard and Kirschner, 1985; Luduena et al., 1988; Lee et al., 1990a,b], while polyglutamylation occurs on residue 438;both residues are within the MAP binding region [Lee et al., 1990a,b; reviewed in Joshi and Cleveland, 19901. These modifications are developmentally regulated and serve to increase the negative charge of the carboxy terminal of class 111. This might be expected to
Differential Beta Tubulin Isotype Sorting
205,
321
-MAP 2
116,
77,
. ,. , EC 02 D4 06
4MAP 2c
__I
a
110* 84,
47,
b.
---EC 02
TAU
0 4 D6
Fig. 6. Western immunoblots of undifferentiated EC cells and neural cells differentiated for 2, 4, and 6 days in retinoic acid. a: Low molecular weight MAP 2C can be faintly detected after 2 days differentiation and continues to increase in expression during days 4 and 6 of RA induced neuronal differentiation. High molecular weight MAP 2 can be detected after 6 days of RA. b: Juvenile tau protein first can be detected by immunoblotting after 6 days RA.
Fig. 5 . MAP 1B expression in EC and in differentiating neural cells. a: In undifferentiated EC cells, MAP 1B co-localizes with MTs in all cells. b: There is increased expression of MAP 1B in a subset of cells after 2 days of neural differentiation. c: MAP 1B is present in neurites, and perhaps also in glial processes, after 6 days of differentiation. Bar, 20 pm.
affect ionic interactions with positively charged MAPs and result in formation of stable MTs high in class I11 beta tubulin and rich in MAPs [Lee et al., 1990a; reviewed by Vallee, 19901. Our analysis shows that during early neuronal differentiation of EC cells, class I1 beta tubulin is more important than class I11 in stable MT formation; however, we do not know when these posttranslational modifications occur in differentiating P 19 cells.
A recent in vitro study demonstrates that bovine brain tubulin, which has been depleted of class I11 beta tubulin, shows decreased polymerization in the presence of either MAP 2 or tau when compared to unfractionated brain tubulin [Banerjee et al., 19901. This would agree with our results which show that class I11 is preferentially localized to the labile MT population as opposed to the stable MT population, and with the results of Joshi and Cleveland [ 19891 which show there is preferential inclusion of class I11 in the soluble fraction and preferential exclusion from the polymeric fraction of differentiating PC12 neurons. It would also explain the diffuse intracellular staining by TuJ1 in our experiments and similar diffuse staining of class I11 beta tubulin in neurons reported by Asai and Remolona [ 19891. It is likely that this diffuse fluorescence, detected in detergent extracted cells, is due to the presence of tubulin oligomers.
322
Falconer et al.
Fig. 7. Muscle cells differentiated for 6 days in the presence of DMSO and double labeled with (a) mouse mAb to muscle-specific actin, MA-931, and with (b) rat mAb YOL 1/34 antibody to visualize all MTs. Muscle cells differentiated for 6 days in DMSO and double
Beta Tubulin lsotypes in Differentiating Muscle Cells
labeled with (c) mouse mAb to class I1 beta tubulin and (d) rat mAb YOL 1/34 to visualize all MTs. Note that essentially all MTs contain class I1 beta tubulin. Bar, 10 pm.
Joshi and Cleveland, 19901; however, less is known about isotype function. Recently, evidence appeared which convincingly demonstrates isotype function Stable MTs in differentiating muscle cells do not [Hoyle and Raff, 19901. In Drosophilu, the functional show selective incorporation of class I1 beta tubulin, de- capabilities of two beta tubulin isoforms are not identispite its high relative percentage in the total MT array. cal. One tubulin isotype, p2, is expressed in postmitotic The expression of class I1 beta tubulin may be unique to germ cells where it is used in all MT functions and armuscle differentiation in EC cultures since only low lev- rays. A second tubulin isotype, p3, is not normally exels of Mp2 (class 11) mRNA expression have been re- pressed in these cells; however when expression is inported in muscle from adult mouse [Lewis and Cowan, duced, in the absence of p2 tubulin, only one MT 1990bl. Differentiating EC muscle cells also show function occurs, spermatid elongation. Axoneme assemsome preferential incorporation of class IV beta tubulin bly, meiosis, and nuclear elongation do not occur, indiinto the stable MT population, which may result from cating that p2 is necessary for these MT functions in interaction with unidentified MAPs or from other, un- these cells. Function(s) for tubulin isotypes in higher known sorting mechanisms. The presence of class I11 vertebrates are still unknown. beta tubulin in muscle samples probably results from the The data presented in this article provide the first presence of some undifferentiated EC cells which do evidence that sub-cellular sorting of two beta tubulin express class I11 at low levels. It does not reflect the isotypes takes place during formation of stable MTs early presence of neurons. in neuronal differentiation. The lack of this sorting in the stable MT population in undifferentiated EC cells and in cells differentiated along the muscle pathway indicates Beta Tubulin lsotype Sorting and Function that this subcellular sorting may be the result of complex Within the past 3 years, several examples of beta interactions between tubulin isotypes and specific tubulin isotype sorting have been identified [reviewed by MAPs.
Differential Beta Tubulin Isotype Sorting
323
ACKNOWLEDGMENTS We thank Drs. D.W. Cleveland, Johns Hopkins University, Baltimore, Maryland; R.F. Luduena, University of Texas, San Antonio; and A. Frankfurter, University of Virginia, Charlottesville, for generously supplying us with antibodies to beta tubulin isotypes, Dr. J.C. Bulinksi and Dr. G.G. Gundersen, Columbia University, New York City for supplying the detyrosinated alpha tubulin antibody, and Dr. L.I. Binder, University of Alabama at Birmingham, Birmingham, Alabama, for the kind gift of antibodies to MAPS lB, 2, and tau. We thank Dr. M. McBurney for the gift of the P19 EC cell line. We would also like to express our thanks to Dr. U. Vielkind for help with the micrographs and to Dr. L.I. Binder, Dr. A. Frankfurter, and Dr. H. Joshi, Emory University, Atlanta, Georgia, for constructive criticism of an earlier version of this manuscript. This research was supported by a NSERC postgraduate fellowship to M.M.F. and by a NSERC grant to D.L.B.
70
60
50 40
30 20
10 REFERENCES
0.
70 60
50 40
30 20
10
0
EC
Neuron
Muscle
Fig. 8. Beta tubulin isotype “profiles” of (a) total MT arrays and (b) stable MT arrays in uncommitted EC cells, neural, and muscle differentiated cells. Mean and standard deviation based on the following number of independent samples: for total MTs-(EC) 5 samples, (neural) 4 samples, (muscle) 3 samples: for stable MTs-(EC) 4 samples, (neural) 3 samples, (muscle) 3 samples.
Aitchison, W.A., and Brown, D.L. (1986): Duplication of the flagellar apparatus and cytoskeletal microtubule system in the alga Polytomella. Cell. Motil. Cytoskeleton 6: 122-127. Arai, T., and Matsumoto, G. (1988): Subcellular localization of functionally differentiated microtubules in squid neurons: regional distribution of microtubule-associated proteins and P-tubulin isotypes. J. Neurochem. 5 1:1825-1 838. Asai, D.J., and Remolona, N.M. (1989): Tubulin isotype usage in vivo: a unique spatial distribution of the minor neuronal-specific P-tubulin isotype in pheochromocytoma cells. Dev. Biol. 132:398-409. Banerjee, A,, Roach, M.C., Wall, K.A., Lopata, M.A., Cleveland, D.W., and Luduena, R.F. (1988): A monoclonal antibody against the type I1 isotype of P-tubulin. J. Biol. Chem. 263: 3029-3034. Banerjee, A., Roach, M.C., Trcka, P., and Luduena, R.F. (1990): Increased microtubule assembly in bovine brain tubulin lacking the type 111 isotype of P-tubulin. J. Biol. Chem. 265:17941799. Black, M.M., and Greene, L.A. (1982): Changes in the colchicine susceptibility of microtubules associated with neurite outgrowth: studies with nerve growth factor-responsive PC 12 pheochromocytoma cells. J. Cell Biol. 95:379-386. Black, M.M., and Keyser, P. (1987): Acetylation of a-tubulin in cultured neurons and the induction of a-tubulin acetylation in PC12 cells by treatment with NGF. J. Neurosci. 7:1833-1842. Black, M.M., Aletta, J.M., and Greene, L.A. (1986): Regulation of microtubule composition and stability during nerve growth factor-promoted neurite outgrowth. J. Cell Biol. 103:545-557. Cadrin, M., Wasteneys, G.O., Jones-Villeneuve, E.M.V., Brown, D.L., and Reuhl, K.R. (1988): Effects of methylmercury on retinoic acid-induced neuroectodermal derivatives of embryonal carcinoma cells. Cell Biol. Toxicol. 4:61-80. Cambray-Deakin, M.A., and Burgoyne, R.D. (1987): Posttranslational modifications of a-tubulin: acetylated and detyrosinated forms in axons of rat cerebellum. J. Cell Biol. 104:1569-1574. Connolly, J.A., and Kalnins, V.I. (1980): Tau and HMW microtu-
324
Falconer et al.
bule-associated proteins have different microtubule binding sites in vivo. Eur. J . Cell Biol. 21:296-300. Denoulet, P., Filliatreau, G., de Nechaud, B., Gros, F., and DiGiamberadino, L. (1989): Differential axonal transport of isotubulins in the motor axons of the rat sciatic nerve. J . Cell Biol. 108: 965 -97 1. Diaz-Nido, J . , and Avila, J . (1989): Quantification of microtubuleassociated protein 1B in brain and other tissues. Int. J . Biochem. 21:723-730. Dinsmore, J . , and Solomon, F. (1991): Ablation of MAP2 expression affects both morphological and cell division phenotypes of neuronal differentiation. Cell 64:817-826. Drubin, D.G., Feinstein, S.C., Shooter, E.M., and Kirschner, M.W. (1985): Nerve growth factor-induced neurite outgrowth in PC 12 cells involves the coordinate induction of microtubule assembly and assembly-promoting factors. J . Cell Biol. 101: 1799-1 807. Edwards, M.K.S., Harris, J.F., and McBurney, M.W. (1983): Induced muscle differentiation in an embryonal carcinoma cell line. Mol. Cell. Biol. 3:2280-2286. Falconer, M.M., Vielkind, U., and Brown, D.L. (1989a): Association of acetylated microtubules, vimentin intermediate filaments and MAP 2 during early neural differentiation in EC cell culture. Biochem. Cell Biol. 67537-544. Falconer, M.M., Vielkind, U . , and Brown, D.L. (1989b): Establishment of a stable, acetylated microtubule bundle during neuronal commitment. Cell Motil. Cytoskeleton 12: 169-180. Ferreira, A., and Caceres, A. (1989a): The expression of acetylated microtubules during axonal and dendritic growth in cerebellar macroneurons which develop in vitro. Dev. Brain Res. 49: 205-2 13. Ferreira, A . , Busciglio, J . , and Caceres, A. (1989b): Microtubule formation and neurite growth in cerebellar macroneurons which develop in vitro: evidence for the involvement of the microtubule-associated proteins, MAP- 1 a, HMW-MAP2 and Tau. Dev. Brain Res. 49:215-228. Fulton, C., and Simpson, P.A. (1976): Selective synthesis and utilization of flagellar tubulin. The multitubulin hypothesis. In Eds. Goldman, R., Pollard, T., and Rosenbaum, J . (eds.): “Cell Motility.” New York: Cold Spring Harbor Publications, pp. 987-1005. Gard, D.L., and Kirschner, M.W. (1985): A polymer-dependent increase in phosphorylation of P-tubulin accompanies differentiation of a mouse neuroblastoma cell line. J. Cell Biol. 100: 764-774. Garner, C.G., Garner, A., Huber, G., Kozak, C., and Matus, A. (1990): Molecular cloning of microtubule-associated protein 1 (MAP1 A) and microtubule-associated protein 5 (MAPlB): identification of distinct genes and their differential expression in developing brain. J. Neurochem. 55:146-154. Gundersen, G.G., Kalnoski, M.H., and Bulinski, J.C. (1984): Distinct populations of microtubules: tyrosinated and nontyrosinated alpha tubulin are distributed differently. Cell 38:779-789. Hoffman, P.N., and Cleveland, D.W. (1988): Neurofilament and tubulin expression recapitulates the developmental program during axonal regeneration: induction of a specific P-tubulin isotype. Proc. Natl. Acad. Sci. U.S.A. 85:4530-4533. Hoyle, H.E., and Raff, E.C. (1990): Two Drosophila beta tubulin isoforms are not functionally equivalent. J . Cell Biol. 111: 1009- 1026. Jones-Villeneuve, E.M.V., McBurney, M.W., Rogers, K.A., and Kalnins, V.I. (1982): Retinoic acid induces embryonal carcinoma cells to differentiate into neurons and glial cells. J . Cell Biol. 94:253 -262.
Joshi, H.C., and Cleveland, D.W. (1989): Differential utilization of P-tubulin isotypes in differentiating neurites. J . Cell Biol. 109: 663-673. Joshi, H.C., and Cleveland, D.W. (1990): Diversity among tubulin subunits: toward what functional end? Cell Motil. Cytoskeleton 16:159-163. Kanai, Y., Takemura, R., Oashima, T., Mori, H., Ihara, Y., Yanagisawa, M., Masaki, T., and Hirokawa, N. (1989): Expression of multiple tau isofonns and microtubule bundle formation in fibroblasts transfected with a single tau cDNA. J. Cell Biol. 10911173-1 184. Khawaja, S . , Gundersen, G.C., and Bulinski, J.C. (1988): Enhanced stability of microtubules enriched in detyrosinated tubulin is not a direct function of detyrosination level. J . Cell Biol. 106: 141-149. Kilmartin, J.V., Wright, B., and Milstein, C. (1982): Rat monoclonal antitubulin antibodies derived by using a new nonsecreting rat cell line. J . Cell Biol. 93:576-582. Kim, H., Jensen, C.G., and Rebhun, L.I. (1986): The binding of MAP 2 and tau on brain microtubules in vitro: implications for microtubule structure. Ann. N.Y. Acad. Sci. 466: 21 8-239. Lee, M.K., Rebhun, L.I., and Frankfurter, A. (1990a): Posttranslational modification of class 111 P-tubulin. Proc. Natl. Acad. Sci. U.S.A. 87:7195-7199. Lee, M.K., Tuttle, J.B., Rebhun, L.I., Cleveland, D.W., and Frankfurter, A. (1990b): The expression and posttranslational modification of a neuron-specific P-tubulin isotype during chick embryogenesis. Cell Motil. Cytoskeleton 17: 118-132. Lewis, S.A., and Cowan, N.J. (1990a): Microtubule bundling. Nature 345:674. Lewis, S.A., and Cowan, N.J. (1990b): Tubulin genes: structure, expression, and regulation. In Avila, J . (ed): “Microtubule Proteins.” Boca Raton, FL: CRC Press, pp. 37-66. Lewis, S.A., Gu, W., and Cowan, N.J. (1987): Free intermingling of mammalian P-tubulin isotypes among functionally distinct microtubules. Cell 49539-548. Lewis, S.A., Ivanov, I.E., Lee, G.-H., and Cowan, N.J. (1989): Organization of microtubules in dendrites and axons is determined by a short hydrophobic zipper in microtubule-associated proteins MAP 2 and tau. Nature 342:498-505. Lim, S-S., Sammak, P.J., and Borisy, G.G. (1989): Progressive and spatially differentiated stability of microtubules in developing neuronal cells. J . Cell Biol. 109:253-263. Littauer, U.Z., Giveon, D., Thierauf, M., Ginzburg, I., and Ponstingl, H. (1986): Common and distinct tubulin binding sites for microtubule-associated proteins. Proc. Natl. Acad. Sci. U. S. A. 83:7 162-7 166. Lopata, M.S., and Cleveland, D.W. (1987): In vivo microtubules are copolymers of available P-tubulin isotypes: localization of each of six vertebrate P-tubulin isotypes using polyclonal antibodies elicited by synthetic peptide antigens. J . Cell Biol. 105:17071720. Luduena, R.F., Zimmerman, H., and Little, M. (1988): Identification of the phosphorylated P-tubulin isotype in differentiated neuroblastoma cells. FEBS Lett. 230:142-146. McBurney, M.W., and Rogers, B.J. (1982): Isolation of male embryonal carcinoma cells and their chromosome replication patterns. Dev. Biol. 89:503-508. McBurney, M.W., Reuhl, K.R., Ally, A.I., Nasipuri, S . , Bell, J.C., and Craig, J. (1988): Differentiation and maturation of embryonal carcinoma-derived neurons in cell culture. J . Neurosci. 8:1063-1073. Murphy, D.B. (1988): Tubulin subunit sorting: analysis of the mech-
Differential Beta Tubulin Isotype Sorting anisms involved in the segregation of unique tubulin isotypes within microtubule copolymers. Protoplasma 145:176-181. Riederer, B., Cohen, R., and Matus, A. (1986): MAPS: a novel brain microtubule-associated protein under strong developmental regulation. J. Neurocytol. 15:763-775. Rudnicki, M.A., Sawtell, N.M., Reuhl, K.R., Berg, R., Craig, J.C., Jardine, K., Lessard, J.L., and McBurney, M.W. (1990): Smooth muscle actin expression during P19 embryonal carcinoma differentiation in cell culture. J. Cell. Physiol. 142:8998. Schulze, E., Asai, D.J., Bulinski, J.C., and Kirschner, M. (1987): Posttranslational modification and microtubule stability. J. Cell Biol. 105:2 167-2 177. Sullivan, K.F. (1988): Structure and utilization of tubulin isotypes. Annu. Rev. Cell Biol. 4:687-716.
325
Tsukada, T., Tippins, D., Gordon, D., Ross, R., and Gown, A.M. (1987): HHF35, a muscle-actin-specific monoclonal antibody. Am. J. Pathol. 126:51-60. Tucker, R.P., and Matus, A.I. (1988): Microtubule-associated proteins characteristic of embryonic brain are found in the adult mammalian retina. Dev. Biol. 130:423-434. Vallee, R.B. (1990): Molecular characterization of high molecular weight microtubule-associated proteins: some answers, many questions. Cell Motil. Cytoskeleton 15:204-209. Wasteneys, G.O., Cadrin, M., Reuhl, K.R., and Brown, D.L. (1988): The effects of methylmercury on the cytoskeleton of murine embryonal carcinoma cells. Cell Biol. Toxicol. 4:41-60. Webster, D.R., Wehland, J., Weber, K., and Borisy, G.G. (1990): Detyrosination of alpha tubulin does not stabilize microtubules in vivo. J. Cell Biol. 111:113-122.