Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Expression of Amino- and Carboxyl-terminal γ- and α-Tubulin Mutants in Cultured Epithelial Cells

Expression of Amino- and Carboxyl-terminal γ- and α-Tubulin Mutants in Cultured Epithelial Cells THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 273, No. 5, Issue of January 30, pp. 2661–2668, 1998 © 1998 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Expression of Amino- and Carboxyl-terminal g- and a-Tubulin Mutants in Cultured Epithelial Cells* (Received for publication, October 1, 1997, and November 6, 1997) Andrew Leask‡§¶ and Tim Stearns§ From ‡FibroGen, Inc., South San Francisco, California 94080-6902 and the §Department of Biological Sciences, Stanford University, Stanford, California 94030-5020 Three distinct tubulin proteins are essential for mi- whereas the growing (plus) end extends away from the MTOC a-, b-, and g-tubulin. After transla- crotubule function: (2). a- and b-tubulin proteins combine into a soluble, 7 tion, De novo microtubule synthesis (that is, synthesis from solu- S heterodimer that is multimerized to form the microtu- ble7S a- and b-heterodimers) occurs at the MTOC. In the past g-tubulin combines with sev- bule filament. Conversely, few years, much information has been gathered as to how this eral proteins into a soluble, 25 S multi-protein particle, process occurs. In cycling animal cells, the main MTOC is the the gammasome that is essential for nucleating micro- centrosome, a nucleus-associated organelle that consists of cen- tubule filaments at the centrosome. The proteins that trioles and pericentriolar material. Among the many centroso- assist tubulins in executing their specific functions are mal proteins is a member of the tubulin family, g-tubulin (3, 4). largely unknown. As an initial approach to address this This protein is evolutionarily conserved, occurring in animals, issue, we first decided to identify domains of mamma- insects, plants, and fungi (3, 5–13). g-Tubulin is not an obligate a- and g-tubulin necessary for their function by lian g-tubulin to centrosomal component (14, 15); yet recruitment of a- and g-tubulin (both de- creating mutant mammalian the centrosome is essential for microtubule nucleation in vivo letion and hybrid mutants) and assaying their behavior and in vitro (5, 8, 9, 11, 14, 16). That the sequence and subcel- in stably transfected Chinese hamster ovary epithelial lular localization of g-tubulin are evolutionarily conserved and cells. First, we demonstrated that addition of a carboxyl- that g-tubulin is essential for microtubule nucleation suggest a terminal epitope tag had no effect on the subcellular a- and g-tubulin. Second, we found localization of either universal mechanism in which g-tubulin plays an essential, g-tubulin that both the amino and carboxyl termini of universal role at the MTOC, possibly by nucleating microtu- were essential for its incorporation into the gamma- bule growth by directly interacting with tubulin heterodimers. some. Third, we found that the amino and carboxyl ter- In cells, cytosolic (that is, soluble, noncentrosomal) g-tubulin a-tubulin were necessary for incorporation of mini of is always found as a large 25 S, ring-like complex (termed a a-b-tubulin heterodimer into the microtubule fila- the gammasome or gTuRC) that nucleates microtubule assembly a-tubulin sequences could not ment network. In general, when recruited to the centrosome (13, 17, 18). That g-tubulin is g-tubulin and vice versa. Taken to- replace those of always found in the gammasome has led to the inference that gether, these results suggest that the amino and car- g-tubulin must interact with other gammasomal proteins to a- and g-tubulin and perhaps regions boxyl termini of exert its function. Although some proteins that might aid g-tu- throughout these proteins were necessary for their spe- bulin in its role have been identified, no direct evidence has cific functions. been provided proving their function. For example, the gam- masome includes a- and b-tubulin (13). Also, an additional protein that may interact with g-tubulin at the centrosome has Microtubule filaments are essential components of the eu- been recently identified (19). However, exactly how g-tubulin karyotic cytoskeleton. For example, they distribute organelles interacts with other proteins to become incorporated into the and mRNAs to various subcellular locales, form the mitotic centrosome is unknown. spindle, and provide the structure for cilia, flagella, and axons. In cells, 7 S tubulin heterodimers are added to the growing The microtubule filament consists of a hollow cylinder com- filament at the centrosome. That growing microtubules usually posed of 13 protofilaments. These protofilaments are in turn have long, curved sheet-like extensions at their plus end has composed of many a- and b-heterodimers arranged in a head to led to the belief that soluble, 7 S a- and b-heterodimers are tail fashion. In most cells, the number, orientation, and nucle- added to the free end of two-dimensional sheets whose lateral ation of microtubules are controlled by a discrete structure, the ends eventually circularize to form a cylinder (20) such that microtubule organizing center (MTOC) (1). The nongrowing a-tubulin subunits are adjacent to a-tubulin subunits of an- (minus) end of the filaments remains attached to the MTOC, other protofilament (21–24). However, exactly how the a-b- heterodimer interacts with other proteins to become stably incorporated into the growing filament is unknown. * This work was supported by funds from the Medical Research Council of Canada (to A. L.) and the National Institutes of Health (to Tubulin proteins show sequence similarity, yet are targetted T. S.). The costs of publication of this article were defrayed in part by to different subcellular localizations and possess different func- the payment of page charges. This article must therefore be hereby tions. The basis for these differences must lie at the level of marked “advertisement” in accordance with 18 U.S.C. Section 1734 amino acid sequence, which must specify the protein-protein solely to indicate this fact. To whom correspondence should be addressed: FibroGen, Inc., 260 interactions necessary for executing tubulin function. Through- Littlefield Ave., S. San Francisco, CA 94080-6902. Tel.: 650-635-1500; out evolution, each tubulin type shares approximately 65–70% Fax: 650-635-1512; E-mail: [email protected]. sequence conservation (25, 26). On the other hand a-, b-, and The abbreviations used are: MTOC, microtubule organizing center; g-tubulin share only about 30% sequence identity (25). Be- CHO, Chinese hamster ovary; CMV, cytomegalovirus; PBS, phosphate- buffered saline. tween tubulins, there is considerably more sequence similarity This paper is available on line at http://www.jbc.org 2661 This is an Open Access article under the CC BY license. 2662 Behavior of Tubulin Mutants in CHO Cells FIG.1. Schematic diagram of basic plasmid used in our transfection experiments. Versions of human g- and mouse a-tubulin genes, both mutant and wild-type, were generated by polymerase chain reaction as described under “Materials and Methods” and were subcloned into the vector HindIII and SalI restriction endonucleases. Genes were subcloned into pTS335 (see “Materials and Methods”) downstream of the strong viral CMV promoter and upstream in-frame to the myc epitope tag, enabling detection of the transfected protein by Western blot and immuno- cytochemical analyses. Line with arrow, CMV promoter; H3, HindIII restriction endonuclease site; Sal, SalI restriction endonuclease site; black box, myc tag; gray box, G418 resistance gene; striped box, kanamycin resistance gene; open box, subcloned gene of interest. ing cell pellets in 8 M urea. SDS-polyacrylamide gel electrophoresis was at the amino termini than at the carboxyl termini. This obser- performed as described using 8.5% polyacrylamide gels (27). Gels were vation has led to the hypothesis that whereas the amino ter- blotted onto nitrocellulose and blocked overnight at 4 °C in 5% nonfat minus of tubulins contains information that pertains to a dry milk, 1 3 PBS, 0.1% Tween 20. Blots were incubated for1hat room shared function among tubulins, the carboxyl terminus con- temperature with each antibody, diluted in 1% milk, 1 3 PBS, 0.1% tains information that is specific to the function of a particular Tween 20. Proteins were detected using a chemiluminescent kit as tubulin type (25). Clearly, tubulins play an essential role in described by the manufacturer (Renaissance, NEN Life Science Prod- ucts). Antibodies used at 1:1000 dilution were: anti-g-tubulin (XGC1– 4; microtubule polymerization and maintenance. To gain new Ref. 14); anti-a-tubulin (YL1/2); anti-b-tubulin, and anti-myc. Second- insights into tubulin function, we decided to determine the ary antibodies (used at a 1:10,000 dilution) were conjugated with horse- roles of the amino and carboxyl termini of a- and g-tubulin in radish peroxidase (Jackson). their subcellular localizations and microtubule filament forma- Gel Filtration and Sucrose Gradient Analysis—Cell extracts for gel tion. To address this issue, we expressed epitope-tagged a- and filtration were prepared by hypotonic lysis. Briefly, cell pellets were g-tubulin mutants in permanently transfected Chinese ham- washed twice in PBS and incubated in 10 mM NaCl, 50 mM Tris-HCl, 2 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, pH 7.5, at 4 °C, fol- ster ovary K1 (CHO-K1) epithelial cell lines. Using immuno- lowed by Dounce homogenization. Extracts were subsequently adjusted fluorescence microscopy, gel filtration chromatography, and to 1 M NaCl. Gel filtration was performed at 4 °C using a Pharmacia fast sucrose gradient centrifugation, we have assayed the cell bio- performance liquid chromatograph with a Superdex 200 column (Phar- logical and biochemical properties of these mutants and have macia Biotech Inc.), with a flow rate of 0.25 ml/min. To verify molecular identified regions necessary for microtubule polymerization. masses corresponding to peaks coming off the column, fast performance liquid chromatography molecular mass standards (Pharmacia) were MATERIALS AND METHODS run with each experiment. Columns were run and loaded in 1 M NaCl, DNA Subcloning—DNA constructs were made using the polymerase 50 mM Tris-HCl, 2 mM EDTA, pH 7.5. Fractions (0.5 ml) were collected chain reaction with vent polymerase (New England Biolabs, Beverly, and concentrated using Microcon 10 concentrators (Amicon, Beverly, MA) and purchased DNA primers (Life Technologies, Inc.) as described MA). SDS sample buffer was added to the samples, which were ana- (27). Primers were designed such that that fragment could be cut with lyzed using SDS/polyacrylamide gel electrophoresis. HindIII and SalI (New England Biolabs) and be subcloned into the Extracts for sucrose gradients were performed by lysing cells in 150 HindIII and SalI sites of the vector pTS335 (a derivative of pCDNA I mM NaCl, 50 mM Tris-HCl, 2 mM EDTA, 2 mM phenylmethylsulfonyl Neo; Invitrogen, San Diego, CA). In the case of mutants in which the fluoride, 0.1% Triton X-100, pH 7.5. Sucrose gradient (10 – 40%) analy- amino-terminal thirds were switched, primers were generated that sis was performed in buffer with 150 mg of extract as described (14). contained a silent mutation that introduced a BstEII site. Thus, hybrids Fractions were collected (either 100 or 150 ml) and processed by SDS/ were generated by a three-piece HindIII/BstEII/SalI ligation. Frag- polyacrylamide gel electrophoresis. Molecular mass standards (Phar- ments were subcloned upstream in-frame to the myc epitope tag and macia) were run in parallel to each gradient. downstream of the strong viral CMV promoter, enabling expression of a RESULTS carboxyl-terminal myc-tagged protein. Correct sequence of the con- structs was verified using Sequenase (U. S. Biochemical Corp.) as de- myc-tagged Human g-Tubulin Is Correctly Localized to the scribed by the manufacturer. Centrosome—Previously, we had demonstrated the importance Cell Culture and Transfection—CHO-K1 cells (American Type Cul- of g-tubulin in initiating microtubule assembly at centrosomes ture Collection, Rockville, MD) were cultured in F-12 medium supple- (14). To assess how g-tubulin is selected to be localized to the mented with 10% fetal calf serum (Life Technologies, Inc.). Plasmids were prepared as described by the manufacturer (Qiagen, Santa centrosome, we subcloned full-length human g-tubulin up- Clarita, CA). Cells were transfected with Lipofectin as described by the stream in-frame to the myc epitope tag and downstream of the manufacturer (Life Technologies, Inc.), and stable transfectants were strong viral CMV promoter (Ref. 28; Fig. 1). Introduction of the selected for and maintained in 800 mg/ml G418 (Life Technologies, Inc.). myc tag at the carboxyl terminus ensured that the introduced, Immunofluorescence Microscopy—Cells were grown on glass cover- mutant protein could be detected by immunofluorescence and slips (VWR, San Francisco, CA). Cells on coverslips were fixed in Western blotting by using an anti-myc antibody. Cell extracts 220 °C methanol for 20 min, washed twice in PBS, and blocked at room temperature for 30 min in PBS that contained 3% bovine serum albu- from cell lines were subjected to Western blotting to verify min and 0.1% Triton X-100. Primary antibodies were then added to expression of the introduced protein and to verify that the fresh solution and incubated for 30 min. Polyclonal antibodies used (at protein could be detected before cells were processed for immu- 1:300 dilution) were: anti-g-tubulin (XGC1– 4, raised against a region in nofluorescence. Thus, upon transfection of constructs into the carboxyl terminus of Xenopus g-tubulin; Ref. 3); anti-a-tubulin mammalian cells, high levels of transgenic protein expression (YL1/2, raised against the entire rat a-tubulin protein; Sera Labs); and was achieved, and the subcellular localization of the introduced anti-b-tubulin (Boehringer Mannheim). The monoclonal anti-myc anti- body was used at 1:300 dilution (9E10). After washing coverslips three protein could be readily detected by immunofluorescent stain- times in PBS, coverslips were incubated in a fresh solution containing ing with an anti-myc antibody. a 1:300 dilution of secondary Texas Red or fluorescein isothiocyanate- Because plasmids used in our study contained a gene confer- conjugated secondary antibodies (Jackson Laboratories, West Grove, ring resistance to neomycin, stably transfected cell lines could PA) for 30 min each before washing and mounting. Nuclei were stained be isolated by growing cells in the drug G418. Possessing per- in 4,6-diamidino-2-phenylindone. Cells were examined using an im- manent cell lines expressing g-tubulin was a necessary prereq- munofluorescence microscope (model Axioskop, Carl Zeiss, Inc., Thornwood, NY). uisite for studying transfected mutant protein because tran- Polyacrylamide Gel Electrophoresis and Immunoblot Analysis—Cell siently transfected g-tubulin was dispersed throughout the extracts were prepared by washing cells twice in PBS and by solubiliz- cytoplasm. CHO-K1 epithelial cells were transfected with a construct encoding full-length myc-tagged g-tubulin. Perma- T. Stearns, unpublished observation. nent cell lines were generated. From these, cell extracts were Behavior of Tubulin Mutants in CHO Cells 2663 FIG.2. Expression of myc-tagged g-tubulin in CHO-K1 cells. Cell extracts (100 mg) were prepared from untransfected CHO-K1 cells and CHO-K1 cells expressing myc-tagged g-tubulin (g-myc) and were subjected to SDS/polyacrylamide gel electrophoresis on 8.5% polyacryl- amide gels. After blotting onto nitrocellulose, blots were probed with anti-myc antibody (A) or anti-g-tubulin antibody (B), as described under “Materials and Methods.” Extracts were from untransfected cells (lanes 1) or cells stably transfected with a construct encoding g-myc (lanes 2). FIG.3. Localization of myc-tagged g-tubulin in CHO-K1 cells. The numbers on the left indicate migration of molecular mass markers, Cells on glass coverslips were processed for immunofluorescence as in kDa. described under “Materials and Methods.” Untransfected CHO-K1 cells were stained with anti-g-tubulin (A) or anti-myc antibodies (B). Cells stably transfected with a construct encoding myc-tagged g-tubulin were stained with anti-g-tubulin (C) or anti-myc antibodies (D). myc-tagged prepared and subjected to polyacrylamide gel electrophoresis. g-tubulin is correctly localized to the centrosome (n 5 300). Gels were blotted onto nitrocellulose and probed with anti-myc and anti-g-tubulin antibodies. Anti-g-tubulin antibody de- tected the endogenous g-tubulin in both untransfected and transfected cell extracts (Fig. 2, lanes 1 and 2). In extracts prepared from stably transfected cells, a protein of slower mo- bility corresponding to myc-tagged g-tubulin was detected with both anti-g-tubulin (Fig. 2, lane 2) and anti-myc antibody (Fig. 2, lane 2), verifying expression of the transgenic protein. The g-tubulin-myc protein was correctly localized to the centrosome of these cells as detected by immunofluorescence staining with an anti-myc antibody; the pattern of staining was indistin- guishable from anti-g-tubulin staining of untransfected cells FIG.4. Expression of myc-tagged g-tubulin deletion mutants in (Fig. 3, A–D). Because epitope-tagged g-tubulin was correctly CHO-K1 cells. Cell extracts were prepared untransfected CHO-K1 cells and from cells expressing ND10 g-tubulin-myc (ND10), CD10 g-tu- localized to the centrosome of stably transfected CHO-K1 cells, bulin-myc (CD10), or CD19 g-tubulin-myc (CD19). Blots were prepared we decided to delineate domains of human g-tubulin essential and processed as described in the legend to Fig. 2. Blots were probed for its subcellular localization by introducing deletions into the with anti-g-tubulin antibody (A) or anti-myc antibody (B). g, endoge- amino and carboxyl termini of g-tubulin and assaying their nous g-tubulin; g-myc, transgenic myc-tagged version of g-tubulin; U, behavior in stably transfected CHO-K1 epithelial cells. untransfected cells. The Amino and Carboxyl Termini of Human g-Tubulin Are Essential—Constructs encoding proteins that lacked the first bulin demonstrated that its 10 amino-terminal amino acids 10 amino-terminal amino acid residues (ND10myc) or the last and its 19 carboxyl-terminal amino acids were essential for 10 or 19 carboxyl-terminal residues (CD10myc or CD19myc, centrosomal localization. respectively) were introduced into CHO-K1 cells (see Fig. 6). The Amino-terminal Third and the 19 Carboxyl-terminal Stable cell lines expressing these transgenes were generated as Amino Acids of g-Tubulin Cannot Be Replaced with Those of detected by immunoblot analysis with anti-g-tubulin and anti- a-Tubulin—The impact of deletions on g-tubulin function could myc antibodies (Fig. 4). To determine if these mutant epitope- have been due to the removal of key amino acid residues di- tagged versions of human g-tubulin were correctly localized in rectly responsible for g-tubulin-specific function, for example, CHO-K1 cells, we performed indirect immunofluorescence on by binding to specific proteins. Alternatively, these deletions cells expressing these proteins. Staining cells with an anti-myc may have affected folding of the g-tubulin protein. To differen- antibody showed that although a mutant g-tubulin protein tiate between these two possibilities, we took advantage of the lacking its 10 carboxyl-terminal amino acids (CD10myc) was fact that a- and g-tubulin show about 30% sequence similarity correctly localized to the centrosome (Fig. 5, C and D), g-tubu- at the amino acid level and are presumed to have similar lin proteins lacking their 10 amino-terminal amino acids overall structures (Fig. 6; Ref. 25). Yet, the amino- and carbox- (ND10myc) or their 19 carboxyl-terminal amino acids yl-terminal amino acids of a- and g-tubulin show identity in 4 (CD19myc) were not correctly localized (Fig. 5, A, B, E, and F). of 10 residues at the amino terminus and 2 of 19 residues at the Stable cell lines were generated that expressed for g-tubulin carboxyl terminus. Thus, replacing g-tubulin sequences with proteins that possessed more severe amino- and carboxyl-ter- those of a-tubulin would not be expected to alter overall protein minal deletions. None of these mutant proteins were correctly structure, but if residues in the amino and carboxyl termini of localized (not shown). Thus, deletion analysis of human g-tu- g-tubulin are essential for g-tubulin-specific function, then 2664 Behavior of Tubulin Mutants in CHO Cells FIG.6. Comparison of the amino acid residues of human g-tu- bulin (27) and mouse a-tubulin (35). Residues conserved between g- and a-tubulin are denoted with a dash. An evolutionarily conserved residue (an insertion relative to a-tubulin) in the amino terminus of g-tubulin is in bold (see text). End points of amino-terminal (ND10) and carboxyl-terminal (CD10 and CD19) g-tubulin mutants mentioned in text are shown. Residues that were exchanged between a- and g-tubulin to generate hybrids in which protein thirds were exchanged are denoted by a box. FIG.5. Localization of myc-tagged g-tubulin deletion mutants in CHO-K1 cells. A and B,ND10 g-tubulin-myc. C and D,CD10 g-tubulin-myc. E and F,CD19 g-tubulin-myc. Antibodies employed in the study were anti-g-tubulin (A, C, and E) or anti-myc antibodies (B, D, FIG.7. Summary of behavior of mutant g-tubulins in stably and F). Cells were processed as described in Fig. 3. Neither ND10-g- transfected CHO-K1 cells. White box, g-tubulin sequence; black box, tubulin-myc nor CD19 g-tubulin-myc were localized to the centrosome a-tubulin sequence. Cells were fixed and stained with anti-myc anti- (n 5 300). body, anti-g-tubulin antibody, and, when appropriate, anti-a-tubulin antibody. Whether the myc epitope was present at centrosome is noted (n 5 300). No staining of the microtubule network was observed. g-myc these substitutions might be expected to abrogate localization is wild-type myc-tagged g-tubulin; ND10myc, CD10myc, and CD19myc of g-tubulin to the centrosome. To assess the impact of these were previously described versions of g-tubulin protein. gN10amyc is a sequence differences on tubulin function, we generated a series version of g-tubulin with its 10 amino-terminal amino acid residues of hybrid constructs between mouse a- and human g-tubulin replaced with those of a-tubulin; similarly gC19amyc is a version of g-tubulin with its 19 carboxyl-terminal amino acids replaced with those and created stably transfected CHO-K1 cells expressing these of a-tubulin. gNamyc and gCamyc are versions of g-tubulin that have hybrid proteins. We subjected these cells to immunofluores- their amino- and carboxyl-terminal thirds, respectively, replaced with cence with anti-myc, anti-a-tubulin, or anti-g-tubulin antibod- those of a-tubulin. ies. A summary of these results is shown in Fig. 7. We deter- mined that substituting the 10 amino-terminal amino acids of g-tubulin with those of a-tubulin (gN10amyc) did not abolish some (13, 14, 18), which becomes recruited into the centrosome, centrosomal localization. This region of a- and g-tubulin differs where it nucleates filament assembly (18). Thus, the inability in six of ten positions, yet this difference was not sufficient to of our mutant g-tubulin proteins to become appropriately lo- alter centrosomal localization of g-tubulin; that is, these 10 calized could be caused by at least two factors. First, the mu- amino-terminal residues are not required for any detectable tant g-tubulin might be defective in its ability to be recruited g-tubulin-specific function. Conversely, substituting the ami- into the gammasome. Alternatively, the mutant g-tubulin no-terminal third of g-tubulin with those of a-tubulin (gNamyc) might retain its ability to be incorporated into the gammasome, abolished centrosomal localization. Similarly, replacing the but the mutation abolished the ability of the gammasome to be carboxyl-terminal 19 amino acids of g-tubulin with those of associated with the centrosome. To distinguish between these a-tubulin (gC19amyc) resulted in a protein incapable of local- two possibilities, we performed sucrose density gradient cen- ization. Thus, the amino-terminal third and carboxyl-terminal trifugation with cell extracts prepared from CHO-K1 cells ex- 19 amino acids of g-tubulin are essential for a g-tubulin-specific pressing myc-tagged wild-type g-tubulin (g-myc) or cells ex- function, namely centrosomal localization. pressing our amino- or carboxyl-terminal localization-defective The Amino and Carboxyl Termini of Human g-Tubulin Are (gNamyc or gC19amyc, respectively) hybrid proteins. Fractions Essential for Its Incorporation into the Gammasome (gTuRC)— from the gradients were subjected to Western blot analysis and After translation, g-tubulin is incorporated into the gamma- probed with anti-g-tubulin or anti-myc antibodies. In untrans- Behavior of Tubulin Mutants in CHO Cells 2665 FIG.9. Size of myc-tagged g-tubulin mutant-containing com- plexes in stably transfected CHO-K1 cells: carboxyl-terminal mutant. Sucrose gradient centrifugation was performed, and fractions were collected and assayed as described in the legend to Fig. 8. Extracts were prepared from CHO-K1 cells expressing gC19amyc, a protein in which the 19 carboxyl-terminal amino acids of g-tubulin are replaced with those of a-tubulin. Nitrocellulose filters were probed with anti-myc antibody or anti-g-tubulin antibody as denoted. Lane numbers corre- spond to fraction numbers from top to bottom of the gradient. gC19amyc is not localized to the gammasome. gC19amyc, transgenic gC19amyc protein; *, transgenic protein from which the myc tag has been clipped. with g-tubulin, we decided to determine if amino- and carboxyl- terminal g-tubulin sequences could replace the corresponding sequences from mouse a-tubulin. To initiate this study, we generated CHO-K1 cell lines stably expressing wild type myc- tagged a-tubulin. This protein was correctly integrated into the microtubule network as determined by immunofluorescence with anti-myc antibody (Fig. 10, A and B). Constructs were generated that permitted expression of mutant versions of a-tubulin in transfected cells. Proteins derived from a-tubulin FIG.8. Size of myc-tagged g-tubulin mutant-containing com- were expressed that either had its first 10 amino-terminal plexes in stably transfected CHO-K1 cells: amino-terminal mu- tant. Sucrose gradient centrifugation was performed (“Materials and residues or its 20 carboxyl-terminal residues replaced with Methods” and Ref. 14). Fractions (100 ml) were collected and assayed as those of g-tubulin. Stably transfected CHO-K1 cells expressing described under “Materials and Methods.” Extracts were prepared from the amino-terminal a-tubulin hybrid were generated; the untransfected CHO-K1 cells (top panel) and CHO-K1 cells expressing transgene was located in the cytosol (Fig. 10, C and D). These gNamyc, which has the amino-terminal third of g-tubulin replaced with cells proliferated readily and had no apparent phenotype in that of a-tubulin (middle and bottom panels). Nitrocellulose filters were probed with anti-myc antibody or anti-g-tubulin antibody as denoted. that cells were indistunguishable in size, shape, and doubling Lane numbers correspond to fraction numbers from top to bottom of time to untransfected cells. An additional a-tubulin mutant the gradient. Under the conditions of the assay, fractions 5 and 6 was generated that possessed a proline insertion between correspond to approximately 25 S. gNamyc is not localized to the amino acid residues one and two (see Fig. 6, bold letter). This gammasome. residue is conserved in g-tubulin (14). This mutant protein also fected CHO-K1 cells, g-tubulin was found in a 25 S complex could not integrate into the filament network (not shown). (Fig. 8, top panel; Ref. 14). Conversely, in cells expressing the Collectively, these results suggest that the 10 amino-terminal amino-terminal localization-defective hybrids (gNamyc), the residues of a-tubulin are essential for its appropriate subcellu- mutant protein was not present in the gammasome but was lar localization. instead localized to the first few fractions of the gradient, In contrast to mutations at the amino terminus of a-tubulin, corresponding to a considerably smaller molecule (Fig. 8, mid- no stably transfected CHO-K1 cells expressing a hybrid a-tu- dle and bottom panels). Furthermore, it appeared that the bulin protein that had its carboxyl-terminal 20 amino acids amino-terminal g-tubulin hybrid mutant was cleaved into at replaced with those of g-tubulin could be isolated. Thus, we least one other detectable form corresponding to a carboxyl- hypothesized that expression of the mutant a-tubulin protein terminal fragment of approximately 20 kDa, only detectable caused lethality. To test this notion, we transiently transfected with the anti-myc antibody (Fig. 8, bottom panel). Incubation of CHO-K1 cells with the appropriate DNA construct and exam- the extract with protease inhibitor mixtures had no effect on ined these cells for expression and subcellular localization of protein cleavage (not shown). Similarly, when extracts pre- the carboxyl-terminal a-tubulin hybrid protein. Transient pared from CHO-K1 cells expressing gC19amyc were analyzed, transfection of CHO-K1 cells with a construct encoding for mutant protein was not localized to the gammasome (Fig. 9). wild- type myc-tagged a-tubulin verified that the transfected Collectively, these data suggest that the amino and carboxyl protein correctly localized to the microtubule filament network. termini of g-tubulin were essential for its incorporation into the However, when the carboxyl-terminal a-tubulin hybrid was gammasome and that the corresponding regions of a-tubulin transfected into CHO-K1 cells, no filament network could be could not substitute for these residues. detected with either anti-a-tubulin or anti-myc antibodies; The Amino and Carboxyl Termini of a-Tubulin Are Impor- a-tubulin protein was dispersed throughout the cell. Further- tant for Heterodimer Function—Intrigued by our observations more, some transfected cells expressing the mutant protein 2666 Behavior of Tubulin Mutants in CHO Cells FIG. 10. Localization of myc-tagged mutant a-tubulin proteins in trans- fected CHO-K1 cells. Cells were pro- cessed for immunofluorescent staining as described under “Materials and Meth- ods.” Anti-myc antibody and anti-a-tubu- lin antibodies were used as denoted on cells transfected with wild-type myc- tagged a-tubulin (A and B), a-tubulin that had its 10 amino-terminal amino acids replaced with those of g-tubulin (aN10gmyc) (C and D), and a-tubulin that had its 19 carboxyl-terminal amino acids replaced with those of g-tubulin (aC19gmyc) (E and F). Cells were stained with anti-g-tubulin (A, C, and E) or anti- myc antibodies (B, D, and F). Whereas myc-tagged a-tubulin was correctly local- ized to the microtubule filament network (n 5 300), neither aN10gmyc (n 5 300) nor aC19gmyc (n 5 42) was appropriately localized to the microtubule network but was instead dispersed throughout the cytosol. appeared round (Fig. 10, E and F). Inspection, before fixation, kDa (Fig. 11, top panel). In addition, when extract from of the tissue culture dish containing transfected cells showed CHO-K1 cells stably expressing the amino-terminal a-tubulin an abnormally large number of rounded cells that were readily mutant (aN10gmyc) were passed over our sizing column, detached from the plate upon agitation. aN10gmyc also behaved as a heterodimer (Fig. 11, bottom three The Amino Terminus of Mouse a-Tubulin Is Necessary for panels). These results suggest that although replacing the 10 Incorporation of the Heterodimer into the Microtubule Net- amino-terminal residues of a-tubulin with those of g-tubulin work—Immediately after translation, a-tubulin must be prop- did not abrogate its ability to heterodimerize with b-tubulin, erly folded and incorporated into a distinct soluble form, the 7 the heterodimer bearing the mutation was not able to effec- S a- and b-heterodimer. Thus, the inability of the a-tubulin tively incorporate into the microtubule filament network. mutant hybrid proteins to integrate into the filament network That permanent cell lines expressing the carboxyl-terminal could be because the mutant proteins are incapable of het- a-tubulin hybrid were not isolated and that cells transiently erodimerizing with b-tubulin. Alternatively, the a-tubulin mu- expressing the mutant protein showed collapsed filaments and tant hybrid proteins might be capable of heterodimerization, exhibited the characteristics of dying cells suggested that the but these mutant heterodimers might be incapable of normal mutant tubulin aC19gmyc heterodimerized with b-tubulin and function. To distinguish between these two possibilities, cell integrated into the filament network, where it exerted its effect extracts were prepared from untransfected CHO-K1 cells and in a dominant-negative fashion. Alternatively, the mutant pro- CHO-K1 cells expressing the amino-terminal a-tubulin mutant tein could associate with itself or other proteins, such as b-tu- protein (aN10gmyc). Extracts were prepared and separated by bulin, to form an insoluble aggregate in the cell. However, size over a gel filtration column. Fractions were collected, con- when we passed extracts containing the mutant protein over centrated, and subjected to Western blot analysis. Not surpris- our sizing column, we could not clearly and conclusively distin- ingly, when extracts from untransfected CHO-K1 cells were guish between these two possibilities. Collectively, our results used in our assay, we found that endogenous a-tubulin behaved with a-tubulin suggest that relatively minor alterations in its as a heterodimer, eluting as a molecular mass of around 100 amino or carboxyl termini profoundly affect activity of a-tubu- Behavior of Tubulin Mutants in CHO Cells 2667 that a-tubulin is present on the minus end of the filament (31, 32, 33). Taken together with these studies, our results suggest that the amino terminus of a-tubulin promotes incorporation of the a-b-heterodimer into the filament by interacting with ei- ther g-tubulin at the centrosome or with b-tubulin at the free, plus end of the microtubule. Second, our data suggest that the 20 carboxyl-terminal res- idues of a-tubulin could be essential for maintenance of the microtubule filament. Replacement of these residues with those of g-tubulin resulted in collapse of the microtubule net- work. Transfected cells expressing the mutant protein became rounded and detached from the tissue culture plate. We hy- pothesize that the role of the 20 carboxyl-terminal amino acid residues is to interact with adjacent protofilaments to promote microtubule filament polymerization or stability. The recent determination that, in a filament, adjacent a-tubulin subunits are adjacent to one another (24) suggests that these residues are essential for lateral associations between adjacent a-tubu- lin subunits. An alternative explanation of our data concerning the carboxyl-terminal a-tubulin mutant is that the mutant protein aggregated in the cytosol with itself or with other proteins to form a particle of similar size to the heterodimer and that this product is toxic to the cell. In any event, our data FIG. 11. Size of amino-terminal myc-tagged mutant a-tubulin show that mutating the carboxyl terminus of a-tubulin abol- (aN10gmyc) containing complexes in CHO-K1 cells. Gel filtration ishes its proper function. was performed as described under “Materials and Methods.” Fractions In cells, the gammasome complex is recruited to the centro- (0.5 ml) were collected, concentrated, and subjected to SDS/polyacryl- some; g-tubulin must be present at centrosomes that are capa- amide gel electrophoresis on 8.5% polyacrylamide gels. After blotting on nitrocellulose, filters were probed with anti-myc, anti-b-tubulin, or anti- ble of undergoing microtubule synthesis. However, g-tubulin is a-tubulin antibodies as noted. Extracts were prepared from untrans- not an obligate centrosomal component (14, 15). Thus, for g-tu- fected CHO-K1 cells (top panel) and cells expressing a-tubulin that had bulin to nucleate filament assembly, it must first be integrated its 10 amino-terminal amino acids replaced with those of g-tubulin into the gammasome and then be incorporated into the centro- (three bottom panels). Sizes of molecular mass standards run with each column are shown on top. Arrows denote the transgenic, myc-tagged some. How g-tubulin interacts with other proteins to exert its protein. aN10gmyc behaved as a heterodimer. function is unknown; however, proteins necessary for g-tubulin function are being identified and characterized (19, 34). lin, abolishing its ability to become properly incorporated into Although several g-tubulin mutants have been isolated, their the microtubule filament. phenotype has only been characterized at a gross level, namely by their inability to nucleate filament assembly (8, 9). Until DISCUSSION now, regions of g-tubulin necessary for particular biochemical Microtubules are essential for many cell activities; for exam- interactions have not been characterized. In this report, we ple, they segregate sister chromatids to opposite ends of the cell determined that deleting either the 10 amino-terminal or the and distribute mRNAs and organelles to particular subcellular 19 carboxyl-terminal amino acids of g-tubulin disrupted its locales. Essential to their activity is their inherent polarity, subcellular localization to the gammasome. Similarly, replac- with growing, plus ends emanating from their anchored, cen- ing the carboxyl-terminal 19 amino acids of g-tubulin with trosome-associated minus end. Three tubulins are essential for those of a-tubulin prevented localization to the gammasome. microtubule polymerization: a-, b-, and g-tubulin. a- and b-tu- This result is not surprising given the large sequence diver- bulin form a heterodimer that is the building block of the gence between the two proteins. Conversely, replacing the 10 microtubule filament, whereas g-tubulin can be recruited to the amino-terminal amino acids of g-tubulin with those of a-tubu- centrosome, where it is essential for microtubule nucleation (5, lin had no effect on g-tubulin localization. These results sug- 8, 9, 11, 14, 16). The molecular basis for microtubule polymer- gest that the amino termini of a- and g-tubulin may interact ization has been the subject of much research. with similar proteins. However, substituting the amino-termi- Although much recent work examined how the filament nal third of g-tubulin with that of a-tubulin abolished localiza- grows (for a review, see Ref. 29), relatively little is known about tion of g-tubulin to the centrosome. Thus, replacing the amino- the structural requirements the heterodimer has for its proper terminal third of g-tubulin or its carboxyl-terminal 10 residues integration and recruitment into the microtubule filament. In with those of a-tubulin abolished the ability of g-tubulin to be this report, we discover that two relatively small domains incorporated into gammasome. These results suggest that present in the amino and carboxyl termini of a-tubulin are these regions are essential for a specific g-tubulin function, necessary for heterodimer function. First, our data show that namely its recruitment to the gammasome. In fact, it may be the 10 amino-terminal residues are essential for recruitment of that regions necessary for the centrosomal localization of g-tu- the heterodimer into the filament network; a single amino acid insertion into this region is sufficient to abolish heterodimer bulin are spread out over the entire protein. Previously, it was assumed that the high degree of sequence incorporation into the microtubule filament network. This ob- servation is intriguing in light of the result that beads coated conservation at the amino-terminal end of a-tubulin and g-tu- bulin meant that these regions of the tubulin proteins con- with GTP bound to the plus end of the microtubule (30). Be- cause b-tubulin, but not a-tubulin, has an exchangeable GTP tained information necessary for direct interaction between a- (20), it was concluded that a-tubulin is present at the minus and g-tubulin (25). Conversely, the high degree of sequence end of the filament; that is, it is believed to interact directly divergence between a- and g-tubulins at their carboxyl termini with g-tubulin. Recently, other observations have confirmed was taken as an indication that this region contained tubulin- 2668 Behavior of Tubulin Mutants in CHO Cells 1289 –1301 type-specific information (25). However, our results suggest 10. Oakley, C. E., and Oakley, B. R. (1989) Nature 338, 662– 664 that information necessary for a- and g-tubulin-specific func- 11. Sobel, S. G., and Snyder, M. (1995) J. Cell Biol. 131, 1775–1788 12. Spang, A., Courtney, I., Fackler, U., Matzner, M., and Schiebel, E. (1993) tion is located at both termini and perhaps throughout the J. Cell Biol. 123, 405– 416 length of these proteins, although the sequence requirements 13. Zheng, Y., Wong, M. L., Alberts, B., and Mitchison, T. (1995) Nature 378, seem to be more strict for a-tubulin than for g-tubulin. That is, 578 –583 14. Stearns, T., and Kirschner, M. (1994) Cell 76, 623– 638 the high degree of sequence conservation within any particular 15. Leask, A., Obrietan, K., and Stearns, T. (1997) Neurosci. Lett. 229, 17–20 tubulin gene is an indication of the high degree of specializa- 16. Joshi, H. C., Monica, J. P., McNamara, L., and Cleveland, D. W. (1992) Nature tion resident in each tubulin protein. As further mutations in 356, 80–83 17. Raff, J. W., Kellogg, D. R., and Alberts, B. M. (1993) J. Cell Biol. 121, 823– 835 tubulin genes are isolated and subjected to cytological and 18. Moritz, M., Braunfeld, M. B., Sedat, J. W., Alberts, B., and Agard, D. (1995) biochemical analyses and the corresponding cellular proteins Nature 378, 638 – 640 19. Geissler, S., Pereira, G., Spang, A., Knop, M., Soues, S., Kilmartin, J., and that interact with these domains are characterized, new in- Schiebel, E. (1996) EMBO J. 15, 3899 –3911 sights should be forthcoming into the molecular mechanism 20. Mandelkow, E.-M., Mandelkow, E., and Milligan, R. A. (1991) J. Cell Biol. 114, underlying microtubule initiation, growth, polarity establish- 977–991 21. Amos, L. A., and Klug, A. (1974) J. Cell Sci. 14, 523–549 ment, and function in eukaryotic cells. 22. Wade, R. H., Horowitz, R., and Milligan, R. A. (1995) Proteins Struct. Funct. Genet. 23, 502–509 Acknowledgment—We thank B. Oakley (Ohio State University, 23. Kikkawa, M., Ishakawa, T., Wakabayashi, T., and Hirokawa, N. (1995) Nature Columbus, OH) for human g-tubulin cDNA. 376, 271–274 24. Sosa, H., and Milligan, R. A. (1996) J. Mol. Biol. 260, 742–755 REFERENCES 25. Burns, R. G. (1995) J. Cell Sci. 108, 2123–2130 1. Pickett-Heaps, J. D. (1969) Cytobios 3, 257–280 26. Kube-Granderath, E., and Schliwa, M. (1997) Eur. J. Cell Biol. 72, 287–296 27. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, 2. Bergen, L. G., and Borisy, G. G. (1980) J. Cell Biol. 84, 141–150 J. A., and Struhl, K. (eds) (1987) Current Protocols in Molecular Biology, 3. Stearns, T., Evans, L., and Kirschner, M. (1991) Cell 65, 825– 836 John Wiley & Sons, Inc., New York 4. Moudjou, M., Bordes, N., Paintrand, M., and Bornens, M. (1996) J. Cell Sci. 28. Horio, T., and Oakley, B. R. (1994). J. Cell Biol. 126, 1465–1473 109, 875– 887 29. Wade, R. H., and Hyman, A. A.(1997) Curr. Opin. Cell Biol. 9, 12–17 5. Horio, T., Uzawa, S., Jung, M. K., Oakley, B. R.,Tanaka, K., and Yanagida, M. 30. Mitchison, T. J. (1993) Science 261, 1044 –1047 (1991) J. Cell Sci. 99, 693–700 6. Liu, B., Joshi, H. C., Wilson, T. J., Silflow, C. D., Palevitz, B. A., and Snustad, 31. Kirschner, K., and Mandelkow, E.-M. (1985) EMBO J. 4, 2397–2402 32. Hirose, K., Fan, J., and Amos, L. A. (1995) J. Mol. Biol. 259, 329 –333 D. P. (1990) Plant Cell 6, 303–314 7. Lopez, I., Khan, S., Sevik, M., Cande, W. Z., and Hussey, H. J. (1995) Plant 33. Fan, J., Griffith, A. D., Lockhart, A., and Cross, R. A. (1996) J. Mol. Biol. 259, 325–330 Physiol. 107, 309 –310 8. Marschall, L. G., Jeng R. L., Mulholland J., and Stearns T. (1996) J. Cell Biol. 34. Knop, M., Pereira, G., Geissler, S., Grein, K., and Schiebel, E. (1997) EMBO J. 16, 1550 –1564 134, 443– 454 9. Oakley, B. R., Oakley, C. E., Yoon, Y., and Jung, M. K. (1990) Cell 61, 35. Lewis, S. A., Lee, M. G., and Cowan, N. J. (1985) J. Cell Biol. 101, 852– 861 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry Unpaywall

Expression of Amino- and Carboxyl-terminal γ- and α-Tubulin Mutants in Cultured Epithelial Cells

Leask, Andrew; Stearns, Tim
Journal of Biological Chemistry – Jan 1, 1998
Download PDF
8 pages

Loading next page...
 
/lp/unpaywall/expression-of-amino-and-carboxyl-terminal-and-tubulin-mutants-in-5ap2af89I2

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0021-9258
DOI
10.1074/jbc.273.5.2661
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 273, No. 5, Issue of January 30, pp. 2661–2668, 1998 © 1998 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Expression of Amino- and Carboxyl-terminal g- and a-Tubulin Mutants in Cultured Epithelial Cells* (Received for publication, October 1, 1997, and November 6, 1997) Andrew Leask‡§¶ and Tim Stearns§ From ‡FibroGen, Inc., South San Francisco, California 94080-6902 and the §Department of Biological Sciences, Stanford University, Stanford, California 94030-5020 Three distinct tubulin proteins are essential for mi- whereas the growing (plus) end extends away from the MTOC a-, b-, and g-tubulin. After transla- crotubule function: (2). a- and b-tubulin proteins combine into a soluble, 7 tion, De novo microtubule synthesis (that is, synthesis from solu- S heterodimer that is multimerized to form the microtu- ble7S a- and b-heterodimers) occurs at the MTOC. In the past g-tubulin combines with sev- bule filament. Conversely, few years, much information has been gathered as to how this eral proteins into a soluble, 25 S multi-protein particle, process occurs. In cycling animal cells, the main MTOC is the the gammasome that is essential for nucleating micro- centrosome, a nucleus-associated organelle that consists of cen- tubule filaments at the centrosome. The proteins that trioles and pericentriolar material. Among the many centroso- assist tubulins in executing their specific functions are mal proteins is a member of the tubulin family, g-tubulin (3, 4). largely unknown. As an initial approach to address this This protein is evolutionarily conserved, occurring in animals, issue, we first decided to identify domains of mamma- insects, plants, and fungi (3, 5–13). g-Tubulin is not an obligate a- and g-tubulin necessary for their function by lian g-tubulin to centrosomal component (14, 15); yet recruitment of a- and g-tubulin (both de- creating mutant mammalian the centrosome is essential for microtubule nucleation in vivo letion and hybrid mutants) and assaying their behavior and in vitro (5, 8, 9, 11, 14, 16). That the sequence and subcel- in stably transfected Chinese hamster ovary epithelial lular localization of g-tubulin are evolutionarily conserved and cells. First, we demonstrated that addition of a carboxyl- that g-tubulin is essential for microtubule nucleation suggest a terminal epitope tag had no effect on the subcellular a- and g-tubulin. Second, we found localization of either universal mechanism in which g-tubulin plays an essential, g-tubulin that both the amino and carboxyl termini of universal role at the MTOC, possibly by nucleating microtu- were essential for its incorporation into the gamma- bule growth by directly interacting with tubulin heterodimers. some. Third, we found that the amino and carboxyl ter- In cells, cytosolic (that is, soluble, noncentrosomal) g-tubulin a-tubulin were necessary for incorporation of mini of is always found as a large 25 S, ring-like complex (termed a a-b-tubulin heterodimer into the microtubule fila- the gammasome or gTuRC) that nucleates microtubule assembly a-tubulin sequences could not ment network. In general, when recruited to the centrosome (13, 17, 18). That g-tubulin is g-tubulin and vice versa. Taken to- replace those of always found in the gammasome has led to the inference that gether, these results suggest that the amino and car- g-tubulin must interact with other gammasomal proteins to a- and g-tubulin and perhaps regions boxyl termini of exert its function. Although some proteins that might aid g-tu- throughout these proteins were necessary for their spe- bulin in its role have been identified, no direct evidence has cific functions. been provided proving their function. For example, the gam- masome includes a- and b-tubulin (13). Also, an additional protein that may interact with g-tubulin at the centrosome has Microtubule filaments are essential components of the eu- been recently identified (19). However, exactly how g-tubulin karyotic cytoskeleton. For example, they distribute organelles interacts with other proteins to become incorporated into the and mRNAs to various subcellular locales, form the mitotic centrosome is unknown. spindle, and provide the structure for cilia, flagella, and axons. In cells, 7 S tubulin heterodimers are added to the growing The microtubule filament consists of a hollow cylinder com- filament at the centrosome. That growing microtubules usually posed of 13 protofilaments. These protofilaments are in turn have long, curved sheet-like extensions at their plus end has composed of many a- and b-heterodimers arranged in a head to led to the belief that soluble, 7 S a- and b-heterodimers are tail fashion. In most cells, the number, orientation, and nucle- added to the free end of two-dimensional sheets whose lateral ation of microtubules are controlled by a discrete structure, the ends eventually circularize to form a cylinder (20) such that microtubule organizing center (MTOC) (1). The nongrowing a-tubulin subunits are adjacent to a-tubulin subunits of an- (minus) end of the filaments remains attached to the MTOC, other protofilament (21–24). However, exactly how the a-b- heterodimer interacts with other proteins to become stably incorporated into the growing filament is unknown. * This work was supported by funds from the Medical Research Council of Canada (to A. L.) and the National Institutes of Health (to Tubulin proteins show sequence similarity, yet are targetted T. S.). The costs of publication of this article were defrayed in part by to different subcellular localizations and possess different func- the payment of page charges. This article must therefore be hereby tions. The basis for these differences must lie at the level of marked “advertisement” in accordance with 18 U.S.C. Section 1734 amino acid sequence, which must specify the protein-protein solely to indicate this fact. To whom correspondence should be addressed: FibroGen, Inc., 260 interactions necessary for executing tubulin function. Through- Littlefield Ave., S. San Francisco, CA 94080-6902. Tel.: 650-635-1500; out evolution, each tubulin type shares approximately 65–70% Fax: 650-635-1512; E-mail: [email protected]. sequence conservation (25, 26). On the other hand a-, b-, and The abbreviations used are: MTOC, microtubule organizing center; g-tubulin share only about 30% sequence identity (25). Be- CHO, Chinese hamster ovary; CMV, cytomegalovirus; PBS, phosphate- buffered saline. tween tubulins, there is considerably more sequence similarity This paper is available on line at http://www.jbc.org 2661 This is an Open Access article under the CC BY license. 2662 Behavior of Tubulin Mutants in CHO Cells FIG.1. Schematic diagram of basic plasmid used in our transfection experiments. Versions of human g- and mouse a-tubulin genes, both mutant and wild-type, were generated by polymerase chain reaction as described under “Materials and Methods” and were subcloned into the vector HindIII and SalI restriction endonucleases. Genes were subcloned into pTS335 (see “Materials and Methods”) downstream of the strong viral CMV promoter and upstream in-frame to the myc epitope tag, enabling detection of the transfected protein by Western blot and immuno- cytochemical analyses. Line with arrow, CMV promoter; H3, HindIII restriction endonuclease site; Sal, SalI restriction endonuclease site; black box, myc tag; gray box, G418 resistance gene; striped box, kanamycin resistance gene; open box, subcloned gene of interest. ing cell pellets in 8 M urea. SDS-polyacrylamide gel electrophoresis was at the amino termini than at the carboxyl termini. This obser- performed as described using 8.5% polyacrylamide gels (27). Gels were vation has led to the hypothesis that whereas the amino ter- blotted onto nitrocellulose and blocked overnight at 4 °C in 5% nonfat minus of tubulins contains information that pertains to a dry milk, 1 3 PBS, 0.1% Tween 20. Blots were incubated for1hat room shared function among tubulins, the carboxyl terminus con- temperature with each antibody, diluted in 1% milk, 1 3 PBS, 0.1% tains information that is specific to the function of a particular Tween 20. Proteins were detected using a chemiluminescent kit as tubulin type (25). Clearly, tubulins play an essential role in described by the manufacturer (Renaissance, NEN Life Science Prod- ucts). Antibodies used at 1:1000 dilution were: anti-g-tubulin (XGC1– 4; microtubule polymerization and maintenance. To gain new Ref. 14); anti-a-tubulin (YL1/2); anti-b-tubulin, and anti-myc. Second- insights into tubulin function, we decided to determine the ary antibodies (used at a 1:10,000 dilution) were conjugated with horse- roles of the amino and carboxyl termini of a- and g-tubulin in radish peroxidase (Jackson). their subcellular localizations and microtubule filament forma- Gel Filtration and Sucrose Gradient Analysis—Cell extracts for gel tion. To address this issue, we expressed epitope-tagged a- and filtration were prepared by hypotonic lysis. Briefly, cell pellets were g-tubulin mutants in permanently transfected Chinese ham- washed twice in PBS and incubated in 10 mM NaCl, 50 mM Tris-HCl, 2 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, pH 7.5, at 4 °C, fol- ster ovary K1 (CHO-K1) epithelial cell lines. Using immuno- lowed by Dounce homogenization. Extracts were subsequently adjusted fluorescence microscopy, gel filtration chromatography, and to 1 M NaCl. Gel filtration was performed at 4 °C using a Pharmacia fast sucrose gradient centrifugation, we have assayed the cell bio- performance liquid chromatograph with a Superdex 200 column (Phar- logical and biochemical properties of these mutants and have macia Biotech Inc.), with a flow rate of 0.25 ml/min. To verify molecular identified regions necessary for microtubule polymerization. masses corresponding to peaks coming off the column, fast performance liquid chromatography molecular mass standards (Pharmacia) were MATERIALS AND METHODS run with each experiment. Columns were run and loaded in 1 M NaCl, DNA Subcloning—DNA constructs were made using the polymerase 50 mM Tris-HCl, 2 mM EDTA, pH 7.5. Fractions (0.5 ml) were collected chain reaction with vent polymerase (New England Biolabs, Beverly, and concentrated using Microcon 10 concentrators (Amicon, Beverly, MA) and purchased DNA primers (Life Technologies, Inc.) as described MA). SDS sample buffer was added to the samples, which were ana- (27). Primers were designed such that that fragment could be cut with lyzed using SDS/polyacrylamide gel electrophoresis. HindIII and SalI (New England Biolabs) and be subcloned into the Extracts for sucrose gradients were performed by lysing cells in 150 HindIII and SalI sites of the vector pTS335 (a derivative of pCDNA I mM NaCl, 50 mM Tris-HCl, 2 mM EDTA, 2 mM phenylmethylsulfonyl Neo; Invitrogen, San Diego, CA). In the case of mutants in which the fluoride, 0.1% Triton X-100, pH 7.5. Sucrose gradient (10 – 40%) analy- amino-terminal thirds were switched, primers were generated that sis was performed in buffer with 150 mg of extract as described (14). contained a silent mutation that introduced a BstEII site. Thus, hybrids Fractions were collected (either 100 or 150 ml) and processed by SDS/ were generated by a three-piece HindIII/BstEII/SalI ligation. Frag- polyacrylamide gel electrophoresis. Molecular mass standards (Phar- ments were subcloned upstream in-frame to the myc epitope tag and macia) were run in parallel to each gradient. downstream of the strong viral CMV promoter, enabling expression of a RESULTS carboxyl-terminal myc-tagged protein. Correct sequence of the con- structs was verified using Sequenase (U. S. Biochemical Corp.) as de- myc-tagged Human g-Tubulin Is Correctly Localized to the scribed by the manufacturer. Centrosome—Previously, we had demonstrated the importance Cell Culture and Transfection—CHO-K1 cells (American Type Cul- of g-tubulin in initiating microtubule assembly at centrosomes ture Collection, Rockville, MD) were cultured in F-12 medium supple- (14). To assess how g-tubulin is selected to be localized to the mented with 10% fetal calf serum (Life Technologies, Inc.). Plasmids were prepared as described by the manufacturer (Qiagen, Santa centrosome, we subcloned full-length human g-tubulin up- Clarita, CA). Cells were transfected with Lipofectin as described by the stream in-frame to the myc epitope tag and downstream of the manufacturer (Life Technologies, Inc.), and stable transfectants were strong viral CMV promoter (Ref. 28; Fig. 1). Introduction of the selected for and maintained in 800 mg/ml G418 (Life Technologies, Inc.). myc tag at the carboxyl terminus ensured that the introduced, Immunofluorescence Microscopy—Cells were grown on glass cover- mutant protein could be detected by immunofluorescence and slips (VWR, San Francisco, CA). Cells on coverslips were fixed in Western blotting by using an anti-myc antibody. Cell extracts 220 °C methanol for 20 min, washed twice in PBS, and blocked at room temperature for 30 min in PBS that contained 3% bovine serum albu- from cell lines were subjected to Western blotting to verify min and 0.1% Triton X-100. Primary antibodies were then added to expression of the introduced protein and to verify that the fresh solution and incubated for 30 min. Polyclonal antibodies used (at protein could be detected before cells were processed for immu- 1:300 dilution) were: anti-g-tubulin (XGC1– 4, raised against a region in nofluorescence. Thus, upon transfection of constructs into the carboxyl terminus of Xenopus g-tubulin; Ref. 3); anti-a-tubulin mammalian cells, high levels of transgenic protein expression (YL1/2, raised against the entire rat a-tubulin protein; Sera Labs); and was achieved, and the subcellular localization of the introduced anti-b-tubulin (Boehringer Mannheim). The monoclonal anti-myc anti- body was used at 1:300 dilution (9E10). After washing coverslips three protein could be readily detected by immunofluorescent stain- times in PBS, coverslips were incubated in a fresh solution containing ing with an anti-myc antibody. a 1:300 dilution of secondary Texas Red or fluorescein isothiocyanate- Because plasmids used in our study contained a gene confer- conjugated secondary antibodies (Jackson Laboratories, West Grove, ring resistance to neomycin, stably transfected cell lines could PA) for 30 min each before washing and mounting. Nuclei were stained be isolated by growing cells in the drug G418. Possessing per- in 4,6-diamidino-2-phenylindone. Cells were examined using an im- manent cell lines expressing g-tubulin was a necessary prereq- munofluorescence microscope (model Axioskop, Carl Zeiss, Inc., Thornwood, NY). uisite for studying transfected mutant protein because tran- Polyacrylamide Gel Electrophoresis and Immunoblot Analysis—Cell siently transfected g-tubulin was dispersed throughout the extracts were prepared by washing cells twice in PBS and by solubiliz- cytoplasm. CHO-K1 epithelial cells were transfected with a construct encoding full-length myc-tagged g-tubulin. Perma- T. Stearns, unpublished observation. nent cell lines were generated. From these, cell extracts were Behavior of Tubulin Mutants in CHO Cells 2663 FIG.2. Expression of myc-tagged g-tubulin in CHO-K1 cells. Cell extracts (100 mg) were prepared from untransfected CHO-K1 cells and CHO-K1 cells expressing myc-tagged g-tubulin (g-myc) and were subjected to SDS/polyacrylamide gel electrophoresis on 8.5% polyacryl- amide gels. After blotting onto nitrocellulose, blots were probed with anti-myc antibody (A) or anti-g-tubulin antibody (B), as described under “Materials and Methods.” Extracts were from untransfected cells (lanes 1) or cells stably transfected with a construct encoding g-myc (lanes 2). FIG.3. Localization of myc-tagged g-tubulin in CHO-K1 cells. The numbers on the left indicate migration of molecular mass markers, Cells on glass coverslips were processed for immunofluorescence as in kDa. described under “Materials and Methods.” Untransfected CHO-K1 cells were stained with anti-g-tubulin (A) or anti-myc antibodies (B). Cells stably transfected with a construct encoding myc-tagged g-tubulin were stained with anti-g-tubulin (C) or anti-myc antibodies (D). myc-tagged prepared and subjected to polyacrylamide gel electrophoresis. g-tubulin is correctly localized to the centrosome (n 5 300). Gels were blotted onto nitrocellulose and probed with anti-myc and anti-g-tubulin antibodies. Anti-g-tubulin antibody de- tected the endogenous g-tubulin in both untransfected and transfected cell extracts (Fig. 2, lanes 1 and 2). In extracts prepared from stably transfected cells, a protein of slower mo- bility corresponding to myc-tagged g-tubulin was detected with both anti-g-tubulin (Fig. 2, lane 2) and anti-myc antibody (Fig. 2, lane 2), verifying expression of the transgenic protein. The g-tubulin-myc protein was correctly localized to the centrosome of these cells as detected by immunofluorescence staining with an anti-myc antibody; the pattern of staining was indistin- guishable from anti-g-tubulin staining of untransfected cells FIG.4. Expression of myc-tagged g-tubulin deletion mutants in (Fig. 3, A–D). Because epitope-tagged g-tubulin was correctly CHO-K1 cells. Cell extracts were prepared untransfected CHO-K1 cells and from cells expressing ND10 g-tubulin-myc (ND10), CD10 g-tu- localized to the centrosome of stably transfected CHO-K1 cells, bulin-myc (CD10), or CD19 g-tubulin-myc (CD19). Blots were prepared we decided to delineate domains of human g-tubulin essential and processed as described in the legend to Fig. 2. Blots were probed for its subcellular localization by introducing deletions into the with anti-g-tubulin antibody (A) or anti-myc antibody (B). g, endoge- amino and carboxyl termini of g-tubulin and assaying their nous g-tubulin; g-myc, transgenic myc-tagged version of g-tubulin; U, behavior in stably transfected CHO-K1 epithelial cells. untransfected cells. The Amino and Carboxyl Termini of Human g-Tubulin Are Essential—Constructs encoding proteins that lacked the first bulin demonstrated that its 10 amino-terminal amino acids 10 amino-terminal amino acid residues (ND10myc) or the last and its 19 carboxyl-terminal amino acids were essential for 10 or 19 carboxyl-terminal residues (CD10myc or CD19myc, centrosomal localization. respectively) were introduced into CHO-K1 cells (see Fig. 6). The Amino-terminal Third and the 19 Carboxyl-terminal Stable cell lines expressing these transgenes were generated as Amino Acids of g-Tubulin Cannot Be Replaced with Those of detected by immunoblot analysis with anti-g-tubulin and anti- a-Tubulin—The impact of deletions on g-tubulin function could myc antibodies (Fig. 4). To determine if these mutant epitope- have been due to the removal of key amino acid residues di- tagged versions of human g-tubulin were correctly localized in rectly responsible for g-tubulin-specific function, for example, CHO-K1 cells, we performed indirect immunofluorescence on by binding to specific proteins. Alternatively, these deletions cells expressing these proteins. Staining cells with an anti-myc may have affected folding of the g-tubulin protein. To differen- antibody showed that although a mutant g-tubulin protein tiate between these two possibilities, we took advantage of the lacking its 10 carboxyl-terminal amino acids (CD10myc) was fact that a- and g-tubulin show about 30% sequence similarity correctly localized to the centrosome (Fig. 5, C and D), g-tubu- at the amino acid level and are presumed to have similar lin proteins lacking their 10 amino-terminal amino acids overall structures (Fig. 6; Ref. 25). Yet, the amino- and carbox- (ND10myc) or their 19 carboxyl-terminal amino acids yl-terminal amino acids of a- and g-tubulin show identity in 4 (CD19myc) were not correctly localized (Fig. 5, A, B, E, and F). of 10 residues at the amino terminus and 2 of 19 residues at the Stable cell lines were generated that expressed for g-tubulin carboxyl terminus. Thus, replacing g-tubulin sequences with proteins that possessed more severe amino- and carboxyl-ter- those of a-tubulin would not be expected to alter overall protein minal deletions. None of these mutant proteins were correctly structure, but if residues in the amino and carboxyl termini of localized (not shown). Thus, deletion analysis of human g-tu- g-tubulin are essential for g-tubulin-specific function, then 2664 Behavior of Tubulin Mutants in CHO Cells FIG.6. Comparison of the amino acid residues of human g-tu- bulin (27) and mouse a-tubulin (35). Residues conserved between g- and a-tubulin are denoted with a dash. An evolutionarily conserved residue (an insertion relative to a-tubulin) in the amino terminus of g-tubulin is in bold (see text). End points of amino-terminal (ND10) and carboxyl-terminal (CD10 and CD19) g-tubulin mutants mentioned in text are shown. Residues that were exchanged between a- and g-tubulin to generate hybrids in which protein thirds were exchanged are denoted by a box. FIG.5. Localization of myc-tagged g-tubulin deletion mutants in CHO-K1 cells. A and B,ND10 g-tubulin-myc. C and D,CD10 g-tubulin-myc. E and F,CD19 g-tubulin-myc. Antibodies employed in the study were anti-g-tubulin (A, C, and E) or anti-myc antibodies (B, D, FIG.7. Summary of behavior of mutant g-tubulins in stably and F). Cells were processed as described in Fig. 3. Neither ND10-g- transfected CHO-K1 cells. White box, g-tubulin sequence; black box, tubulin-myc nor CD19 g-tubulin-myc were localized to the centrosome a-tubulin sequence. Cells were fixed and stained with anti-myc anti- (n 5 300). body, anti-g-tubulin antibody, and, when appropriate, anti-a-tubulin antibody. Whether the myc epitope was present at centrosome is noted (n 5 300). No staining of the microtubule network was observed. g-myc these substitutions might be expected to abrogate localization is wild-type myc-tagged g-tubulin; ND10myc, CD10myc, and CD19myc of g-tubulin to the centrosome. To assess the impact of these were previously described versions of g-tubulin protein. gN10amyc is a sequence differences on tubulin function, we generated a series version of g-tubulin with its 10 amino-terminal amino acid residues of hybrid constructs between mouse a- and human g-tubulin replaced with those of a-tubulin; similarly gC19amyc is a version of g-tubulin with its 19 carboxyl-terminal amino acids replaced with those and created stably transfected CHO-K1 cells expressing these of a-tubulin. gNamyc and gCamyc are versions of g-tubulin that have hybrid proteins. We subjected these cells to immunofluores- their amino- and carboxyl-terminal thirds, respectively, replaced with cence with anti-myc, anti-a-tubulin, or anti-g-tubulin antibod- those of a-tubulin. ies. A summary of these results is shown in Fig. 7. We deter- mined that substituting the 10 amino-terminal amino acids of g-tubulin with those of a-tubulin (gN10amyc) did not abolish some (13, 14, 18), which becomes recruited into the centrosome, centrosomal localization. This region of a- and g-tubulin differs where it nucleates filament assembly (18). Thus, the inability in six of ten positions, yet this difference was not sufficient to of our mutant g-tubulin proteins to become appropriately lo- alter centrosomal localization of g-tubulin; that is, these 10 calized could be caused by at least two factors. First, the mu- amino-terminal residues are not required for any detectable tant g-tubulin might be defective in its ability to be recruited g-tubulin-specific function. Conversely, substituting the ami- into the gammasome. Alternatively, the mutant g-tubulin no-terminal third of g-tubulin with those of a-tubulin (gNamyc) might retain its ability to be incorporated into the gammasome, abolished centrosomal localization. Similarly, replacing the but the mutation abolished the ability of the gammasome to be carboxyl-terminal 19 amino acids of g-tubulin with those of associated with the centrosome. To distinguish between these a-tubulin (gC19amyc) resulted in a protein incapable of local- two possibilities, we performed sucrose density gradient cen- ization. Thus, the amino-terminal third and carboxyl-terminal trifugation with cell extracts prepared from CHO-K1 cells ex- 19 amino acids of g-tubulin are essential for a g-tubulin-specific pressing myc-tagged wild-type g-tubulin (g-myc) or cells ex- function, namely centrosomal localization. pressing our amino- or carboxyl-terminal localization-defective The Amino and Carboxyl Termini of Human g-Tubulin Are (gNamyc or gC19amyc, respectively) hybrid proteins. Fractions Essential for Its Incorporation into the Gammasome (gTuRC)— from the gradients were subjected to Western blot analysis and After translation, g-tubulin is incorporated into the gamma- probed with anti-g-tubulin or anti-myc antibodies. In untrans- Behavior of Tubulin Mutants in CHO Cells 2665 FIG.9. Size of myc-tagged g-tubulin mutant-containing com- plexes in stably transfected CHO-K1 cells: carboxyl-terminal mutant. Sucrose gradient centrifugation was performed, and fractions were collected and assayed as described in the legend to Fig. 8. Extracts were prepared from CHO-K1 cells expressing gC19amyc, a protein in which the 19 carboxyl-terminal amino acids of g-tubulin are replaced with those of a-tubulin. Nitrocellulose filters were probed with anti-myc antibody or anti-g-tubulin antibody as denoted. Lane numbers corre- spond to fraction numbers from top to bottom of the gradient. gC19amyc is not localized to the gammasome. gC19amyc, transgenic gC19amyc protein; *, transgenic protein from which the myc tag has been clipped. with g-tubulin, we decided to determine if amino- and carboxyl- terminal g-tubulin sequences could replace the corresponding sequences from mouse a-tubulin. To initiate this study, we generated CHO-K1 cell lines stably expressing wild type myc- tagged a-tubulin. This protein was correctly integrated into the microtubule network as determined by immunofluorescence with anti-myc antibody (Fig. 10, A and B). Constructs were generated that permitted expression of mutant versions of a-tubulin in transfected cells. Proteins derived from a-tubulin FIG.8. Size of myc-tagged g-tubulin mutant-containing com- were expressed that either had its first 10 amino-terminal plexes in stably transfected CHO-K1 cells: amino-terminal mu- tant. Sucrose gradient centrifugation was performed (“Materials and residues or its 20 carboxyl-terminal residues replaced with Methods” and Ref. 14). Fractions (100 ml) were collected and assayed as those of g-tubulin. Stably transfected CHO-K1 cells expressing described under “Materials and Methods.” Extracts were prepared from the amino-terminal a-tubulin hybrid were generated; the untransfected CHO-K1 cells (top panel) and CHO-K1 cells expressing transgene was located in the cytosol (Fig. 10, C and D). These gNamyc, which has the amino-terminal third of g-tubulin replaced with cells proliferated readily and had no apparent phenotype in that of a-tubulin (middle and bottom panels). Nitrocellulose filters were probed with anti-myc antibody or anti-g-tubulin antibody as denoted. that cells were indistunguishable in size, shape, and doubling Lane numbers correspond to fraction numbers from top to bottom of time to untransfected cells. An additional a-tubulin mutant the gradient. Under the conditions of the assay, fractions 5 and 6 was generated that possessed a proline insertion between correspond to approximately 25 S. gNamyc is not localized to the amino acid residues one and two (see Fig. 6, bold letter). This gammasome. residue is conserved in g-tubulin (14). This mutant protein also fected CHO-K1 cells, g-tubulin was found in a 25 S complex could not integrate into the filament network (not shown). (Fig. 8, top panel; Ref. 14). Conversely, in cells expressing the Collectively, these results suggest that the 10 amino-terminal amino-terminal localization-defective hybrids (gNamyc), the residues of a-tubulin are essential for its appropriate subcellu- mutant protein was not present in the gammasome but was lar localization. instead localized to the first few fractions of the gradient, In contrast to mutations at the amino terminus of a-tubulin, corresponding to a considerably smaller molecule (Fig. 8, mid- no stably transfected CHO-K1 cells expressing a hybrid a-tu- dle and bottom panels). Furthermore, it appeared that the bulin protein that had its carboxyl-terminal 20 amino acids amino-terminal g-tubulin hybrid mutant was cleaved into at replaced with those of g-tubulin could be isolated. Thus, we least one other detectable form corresponding to a carboxyl- hypothesized that expression of the mutant a-tubulin protein terminal fragment of approximately 20 kDa, only detectable caused lethality. To test this notion, we transiently transfected with the anti-myc antibody (Fig. 8, bottom panel). Incubation of CHO-K1 cells with the appropriate DNA construct and exam- the extract with protease inhibitor mixtures had no effect on ined these cells for expression and subcellular localization of protein cleavage (not shown). Similarly, when extracts pre- the carboxyl-terminal a-tubulin hybrid protein. Transient pared from CHO-K1 cells expressing gC19amyc were analyzed, transfection of CHO-K1 cells with a construct encoding for mutant protein was not localized to the gammasome (Fig. 9). wild- type myc-tagged a-tubulin verified that the transfected Collectively, these data suggest that the amino and carboxyl protein correctly localized to the microtubule filament network. termini of g-tubulin were essential for its incorporation into the However, when the carboxyl-terminal a-tubulin hybrid was gammasome and that the corresponding regions of a-tubulin transfected into CHO-K1 cells, no filament network could be could not substitute for these residues. detected with either anti-a-tubulin or anti-myc antibodies; The Amino and Carboxyl Termini of a-Tubulin Are Impor- a-tubulin protein was dispersed throughout the cell. Further- tant for Heterodimer Function—Intrigued by our observations more, some transfected cells expressing the mutant protein 2666 Behavior of Tubulin Mutants in CHO Cells FIG. 10. Localization of myc-tagged mutant a-tubulin proteins in trans- fected CHO-K1 cells. Cells were pro- cessed for immunofluorescent staining as described under “Materials and Meth- ods.” Anti-myc antibody and anti-a-tubu- lin antibodies were used as denoted on cells transfected with wild-type myc- tagged a-tubulin (A and B), a-tubulin that had its 10 amino-terminal amino acids replaced with those of g-tubulin (aN10gmyc) (C and D), and a-tubulin that had its 19 carboxyl-terminal amino acids replaced with those of g-tubulin (aC19gmyc) (E and F). Cells were stained with anti-g-tubulin (A, C, and E) or anti- myc antibodies (B, D, and F). Whereas myc-tagged a-tubulin was correctly local- ized to the microtubule filament network (n 5 300), neither aN10gmyc (n 5 300) nor aC19gmyc (n 5 42) was appropriately localized to the microtubule network but was instead dispersed throughout the cytosol. appeared round (Fig. 10, E and F). Inspection, before fixation, kDa (Fig. 11, top panel). In addition, when extract from of the tissue culture dish containing transfected cells showed CHO-K1 cells stably expressing the amino-terminal a-tubulin an abnormally large number of rounded cells that were readily mutant (aN10gmyc) were passed over our sizing column, detached from the plate upon agitation. aN10gmyc also behaved as a heterodimer (Fig. 11, bottom three The Amino Terminus of Mouse a-Tubulin Is Necessary for panels). These results suggest that although replacing the 10 Incorporation of the Heterodimer into the Microtubule Net- amino-terminal residues of a-tubulin with those of g-tubulin work—Immediately after translation, a-tubulin must be prop- did not abrogate its ability to heterodimerize with b-tubulin, erly folded and incorporated into a distinct soluble form, the 7 the heterodimer bearing the mutation was not able to effec- S a- and b-heterodimer. Thus, the inability of the a-tubulin tively incorporate into the microtubule filament network. mutant hybrid proteins to integrate into the filament network That permanent cell lines expressing the carboxyl-terminal could be because the mutant proteins are incapable of het- a-tubulin hybrid were not isolated and that cells transiently erodimerizing with b-tubulin. Alternatively, the a-tubulin mu- expressing the mutant protein showed collapsed filaments and tant hybrid proteins might be capable of heterodimerization, exhibited the characteristics of dying cells suggested that the but these mutant heterodimers might be incapable of normal mutant tubulin aC19gmyc heterodimerized with b-tubulin and function. To distinguish between these two possibilities, cell integrated into the filament network, where it exerted its effect extracts were prepared from untransfected CHO-K1 cells and in a dominant-negative fashion. Alternatively, the mutant pro- CHO-K1 cells expressing the amino-terminal a-tubulin mutant tein could associate with itself or other proteins, such as b-tu- protein (aN10gmyc). Extracts were prepared and separated by bulin, to form an insoluble aggregate in the cell. However, size over a gel filtration column. Fractions were collected, con- when we passed extracts containing the mutant protein over centrated, and subjected to Western blot analysis. Not surpris- our sizing column, we could not clearly and conclusively distin- ingly, when extracts from untransfected CHO-K1 cells were guish between these two possibilities. Collectively, our results used in our assay, we found that endogenous a-tubulin behaved with a-tubulin suggest that relatively minor alterations in its as a heterodimer, eluting as a molecular mass of around 100 amino or carboxyl termini profoundly affect activity of a-tubu- Behavior of Tubulin Mutants in CHO Cells 2667 that a-tubulin is present on the minus end of the filament (31, 32, 33). Taken together with these studies, our results suggest that the amino terminus of a-tubulin promotes incorporation of the a-b-heterodimer into the filament by interacting with ei- ther g-tubulin at the centrosome or with b-tubulin at the free, plus end of the microtubule. Second, our data suggest that the 20 carboxyl-terminal res- idues of a-tubulin could be essential for maintenance of the microtubule filament. Replacement of these residues with those of g-tubulin resulted in collapse of the microtubule net- work. Transfected cells expressing the mutant protein became rounded and detached from the tissue culture plate. We hy- pothesize that the role of the 20 carboxyl-terminal amino acid residues is to interact with adjacent protofilaments to promote microtubule filament polymerization or stability. The recent determination that, in a filament, adjacent a-tubulin subunits are adjacent to one another (24) suggests that these residues are essential for lateral associations between adjacent a-tubu- lin subunits. An alternative explanation of our data concerning the carboxyl-terminal a-tubulin mutant is that the mutant protein aggregated in the cytosol with itself or with other proteins to form a particle of similar size to the heterodimer and that this product is toxic to the cell. In any event, our data FIG. 11. Size of amino-terminal myc-tagged mutant a-tubulin show that mutating the carboxyl terminus of a-tubulin abol- (aN10gmyc) containing complexes in CHO-K1 cells. Gel filtration ishes its proper function. was performed as described under “Materials and Methods.” Fractions In cells, the gammasome complex is recruited to the centro- (0.5 ml) were collected, concentrated, and subjected to SDS/polyacryl- some; g-tubulin must be present at centrosomes that are capa- amide gel electrophoresis on 8.5% polyacrylamide gels. After blotting on nitrocellulose, filters were probed with anti-myc, anti-b-tubulin, or anti- ble of undergoing microtubule synthesis. However, g-tubulin is a-tubulin antibodies as noted. Extracts were prepared from untrans- not an obligate centrosomal component (14, 15). Thus, for g-tu- fected CHO-K1 cells (top panel) and cells expressing a-tubulin that had bulin to nucleate filament assembly, it must first be integrated its 10 amino-terminal amino acids replaced with those of g-tubulin into the gammasome and then be incorporated into the centro- (three bottom panels). Sizes of molecular mass standards run with each column are shown on top. Arrows denote the transgenic, myc-tagged some. How g-tubulin interacts with other proteins to exert its protein. aN10gmyc behaved as a heterodimer. function is unknown; however, proteins necessary for g-tubulin function are being identified and characterized (19, 34). lin, abolishing its ability to become properly incorporated into Although several g-tubulin mutants have been isolated, their the microtubule filament. phenotype has only been characterized at a gross level, namely by their inability to nucleate filament assembly (8, 9). Until DISCUSSION now, regions of g-tubulin necessary for particular biochemical Microtubules are essential for many cell activities; for exam- interactions have not been characterized. In this report, we ple, they segregate sister chromatids to opposite ends of the cell determined that deleting either the 10 amino-terminal or the and distribute mRNAs and organelles to particular subcellular 19 carboxyl-terminal amino acids of g-tubulin disrupted its locales. Essential to their activity is their inherent polarity, subcellular localization to the gammasome. Similarly, replac- with growing, plus ends emanating from their anchored, cen- ing the carboxyl-terminal 19 amino acids of g-tubulin with trosome-associated minus end. Three tubulins are essential for those of a-tubulin prevented localization to the gammasome. microtubule polymerization: a-, b-, and g-tubulin. a- and b-tu- This result is not surprising given the large sequence diver- bulin form a heterodimer that is the building block of the gence between the two proteins. Conversely, replacing the 10 microtubule filament, whereas g-tubulin can be recruited to the amino-terminal amino acids of g-tubulin with those of a-tubu- centrosome, where it is essential for microtubule nucleation (5, lin had no effect on g-tubulin localization. These results sug- 8, 9, 11, 14, 16). The molecular basis for microtubule polymer- gest that the amino termini of a- and g-tubulin may interact ization has been the subject of much research. with similar proteins. However, substituting the amino-termi- Although much recent work examined how the filament nal third of g-tubulin with that of a-tubulin abolished localiza- grows (for a review, see Ref. 29), relatively little is known about tion of g-tubulin to the centrosome. Thus, replacing the amino- the structural requirements the heterodimer has for its proper terminal third of g-tubulin or its carboxyl-terminal 10 residues integration and recruitment into the microtubule filament. In with those of a-tubulin abolished the ability of g-tubulin to be this report, we discover that two relatively small domains incorporated into gammasome. These results suggest that present in the amino and carboxyl termini of a-tubulin are these regions are essential for a specific g-tubulin function, necessary for heterodimer function. First, our data show that namely its recruitment to the gammasome. In fact, it may be the 10 amino-terminal residues are essential for recruitment of that regions necessary for the centrosomal localization of g-tu- the heterodimer into the filament network; a single amino acid insertion into this region is sufficient to abolish heterodimer bulin are spread out over the entire protein. Previously, it was assumed that the high degree of sequence incorporation into the microtubule filament network. This ob- servation is intriguing in light of the result that beads coated conservation at the amino-terminal end of a-tubulin and g-tu- bulin meant that these regions of the tubulin proteins con- with GTP bound to the plus end of the microtubule (30). Be- cause b-tubulin, but not a-tubulin, has an exchangeable GTP tained information necessary for direct interaction between a- (20), it was concluded that a-tubulin is present at the minus and g-tubulin (25). Conversely, the high degree of sequence end of the filament; that is, it is believed to interact directly divergence between a- and g-tubulins at their carboxyl termini with g-tubulin. Recently, other observations have confirmed was taken as an indication that this region contained tubulin- 2668 Behavior of Tubulin Mutants in CHO Cells 1289 –1301 type-specific information (25). However, our results suggest 10. Oakley, C. E., and Oakley, B. R. (1989) Nature 338, 662– 664 that information necessary for a- and g-tubulin-specific func- 11. Sobel, S. G., and Snyder, M. (1995) J. Cell Biol. 131, 1775–1788 12. Spang, A., Courtney, I., Fackler, U., Matzner, M., and Schiebel, E. (1993) tion is located at both termini and perhaps throughout the J. Cell Biol. 123, 405– 416 length of these proteins, although the sequence requirements 13. Zheng, Y., Wong, M. L., Alberts, B., and Mitchison, T. (1995) Nature 378, seem to be more strict for a-tubulin than for g-tubulin. That is, 578 –583 14. Stearns, T., and Kirschner, M. (1994) Cell 76, 623– 638 the high degree of sequence conservation within any particular 15. Leask, A., Obrietan, K., and Stearns, T. (1997) Neurosci. Lett. 229, 17–20 tubulin gene is an indication of the high degree of specializa- 16. Joshi, H. C., Monica, J. P., McNamara, L., and Cleveland, D. W. (1992) Nature tion resident in each tubulin protein. As further mutations in 356, 80–83 17. Raff, J. W., Kellogg, D. R., and Alberts, B. M. (1993) J. Cell Biol. 121, 823– 835 tubulin genes are isolated and subjected to cytological and 18. Moritz, M., Braunfeld, M. B., Sedat, J. W., Alberts, B., and Agard, D. (1995) biochemical analyses and the corresponding cellular proteins Nature 378, 638 – 640 19. Geissler, S., Pereira, G., Spang, A., Knop, M., Soues, S., Kilmartin, J., and that interact with these domains are characterized, new in- Schiebel, E. (1996) EMBO J. 15, 3899 –3911 sights should be forthcoming into the molecular mechanism 20. Mandelkow, E.-M., Mandelkow, E., and Milligan, R. A. (1991) J. Cell Biol. 114, underlying microtubule initiation, growth, polarity establish- 977–991 21. Amos, L. A., and Klug, A. (1974) J. Cell Sci. 14, 523–549 ment, and function in eukaryotic cells. 22. Wade, R. H., Horowitz, R., and Milligan, R. A. (1995) Proteins Struct. Funct. Genet. 23, 502–509 Acknowledgment—We thank B. Oakley (Ohio State University, 23. Kikkawa, M., Ishakawa, T., Wakabayashi, T., and Hirokawa, N. (1995) Nature Columbus, OH) for human g-tubulin cDNA. 376, 271–274 24. Sosa, H., and Milligan, R. A. (1996) J. Mol. Biol. 260, 742–755 REFERENCES 25. Burns, R. G. (1995) J. Cell Sci. 108, 2123–2130 1. Pickett-Heaps, J. D. (1969) Cytobios 3, 257–280 26. Kube-Granderath, E., and Schliwa, M. (1997) Eur. J. Cell Biol. 72, 287–296 27. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, 2. Bergen, L. G., and Borisy, G. G. (1980) J. Cell Biol. 84, 141–150 J. A., and Struhl, K. (eds) (1987) Current Protocols in Molecular Biology, 3. Stearns, T., Evans, L., and Kirschner, M. (1991) Cell 65, 825– 836 John Wiley & Sons, Inc., New York 4. Moudjou, M., Bordes, N., Paintrand, M., and Bornens, M. (1996) J. Cell Sci. 28. Horio, T., and Oakley, B. R. (1994). J. Cell Biol. 126, 1465–1473 109, 875– 887 29. Wade, R. H., and Hyman, A. A.(1997) Curr. Opin. Cell Biol. 9, 12–17 5. Horio, T., Uzawa, S., Jung, M. K., Oakley, B. R.,Tanaka, K., and Yanagida, M. 30. Mitchison, T. J. (1993) Science 261, 1044 –1047 (1991) J. Cell Sci. 99, 693–700 6. Liu, B., Joshi, H. C., Wilson, T. J., Silflow, C. D., Palevitz, B. A., and Snustad, 31. Kirschner, K., and Mandelkow, E.-M. (1985) EMBO J. 4, 2397–2402 32. Hirose, K., Fan, J., and Amos, L. A. (1995) J. Mol. Biol. 259, 329 –333 D. P. (1990) Plant Cell 6, 303–314 7. Lopez, I., Khan, S., Sevik, M., Cande, W. Z., and Hussey, H. J. (1995) Plant 33. Fan, J., Griffith, A. D., Lockhart, A., and Cross, R. A. (1996) J. Mol. Biol. 259, 325–330 Physiol. 107, 309 –310 8. Marschall, L. G., Jeng R. L., Mulholland J., and Stearns T. (1996) J. Cell Biol. 34. Knop, M., Pereira, G., Geissler, S., Grein, K., and Schiebel, E. (1997) EMBO J. 16, 1550 –1564 134, 443– 454 9. Oakley, B. R., Oakley, C. E., Yoon, Y., and Jung, M. K. (1990) Cell 61, 35. Lewis, S. A., Lee, M. G., and Cowan, N. J. (1985) J. Cell Biol. 101, 852– 861

Journal

Journal of Biological Chemistry – Unpaywall

Published: Jan 1, 1998

There are no references for this article.