Biochemistry 1995,34, 7409-7419
7409
Site-Directed Mutagenesis of Putative GTP-Binding Sites of Yeast P-Tubulin: Evidence That a-, /3-, and y-Tubulins Are Atypical GTPasest Carleton R. Sage,;,§ Cynthia A. Dougherty,i Ashley S . Davis,*$”Roy G. Burns,l Leslie Wilson,; and Kevin W. Farrell*s$ Department of Biological Sciences, University of Califomia, Santa Barbara, California 93106, and Biophysics Section, The Blackett Laboratory, Imperial College of Science, Technology and Medicine, London SW7 2BZ, U.K. Received February 9, 1995; Revised Manuscript Received March 30, 1995@
The exchangeable GTP-binding site on ,&tubulin has been extensively studied, but the primary sequence elements which form the binding site on P-tubulin remain unknown. We have used site-directed mutagenesis of the single ,&tubulin gene of Saccharomyces cerevisiae to test a model for the GTPbinding site on P-tubulin, which was based on sequence comparisons with members of the GTPase superfamily [Stemlicht, H., Yaffe, M. B., & Fan, G. W. (1987) FEBS Lett. 214,226-2351. We analyzed the effects of D295N, N298K, and N298Q mutations in a proposed base-binding motif, 295DAKN298, on tubulin-GTP binding and on nucleotide-binding specificity. We also examined the effects of a D203S mutation in a putative phosphate-binding region, 203DNEA206,on nucleotide binding affinity, on the assembly-dependent tubulin GTPase activity in vitro, and on the dynamic properties of individual “mutant” microtubules in vitro. The effects of the mutations on cell phenotype and on microtubule polymerization in cells were also measured. The results do not support the proposal that the 203DNEA206 and 295NKAD298 motifs are cognate to motifs found in GTPase superfamily members. Instead, the data argue that the primary sequence elements of /3-tubulins that interact with bound nucleotide, and presumably also those of the a- and y-tubulin family members, are different from those of “typical” GTPase superfamily members, such as p2lru,’. The GTPase superfamily should thus be broadened to include not just the typical GTPases that show strong conservation of primary sequence consensus motifs (GxxxxGK, T, DxxG, DxKN) but also “atypical” GTPases, exemplified by the tubulins and other recently identified GTPases, that do not show the consensus motifs of typical GTPases and which also show no obvious primary sequence relationships between themselves. The tubulins and other atypical GTPases thus appear to represent convergent solutions to the GTP-binding and hydrolysis problem. ABSTRACT:
The tubulin family of proteins, the a-, /3-, and y-tubulins, are members of a large family of proteins which bind and hydrolyze guanosine triphosphate (GTP). The a- and ,&tubulins normally associate to form a heterodimer which binds 2 mol of GTP (Jacobs, 1975; Luduena et al., 1977). Cross-linking (Geahlen & Haley; 1977; Maccioni & Seeds, 1983; Nath et al., 1985; Steiner, 1984) and immunological (Hesse et al., 1985) studies have shown that GTP binding to one site (the E-site) is readily exchangeable and is located on /3-tubulin. By contrast, GTP is bound nonexchangeably at the second site (the N-site) and is presumed to be on the a-subunit. Furthermore, only P-tubulin is a bona fide GTPase and hydrolyzes GTP during microtubule formation [reviewed in Dustin (1984)l; GTP hydrolysis by a-tubulin has not been observed, and the nucleotide-binding and hydrolysis properties of y-tubulin are not known. Somewhat unusual, however, is the observation that none of the tubulins show the characteristic arrangement of primary sequence ‘This work was supported by NIH Grants GM41751 (K.W.F.) and NS13560 (L.W.). * To whom correspondence should be addressed [telephone, (805) 893-385 1; FAX. (805) 893-4724]. University of California, Santa Barbara. Present address: Department of Biochemistry and Biophysics, Box 0448, University of Califomia, San Francisco, CA 94143-0448. I’ Present address: Cytoskeleton, 1899 Gaylord St., Denver, CO 80206. Imperial College of Science, Technology and Medicine. Abstract published in Advance ACS Abstracts, May 15, 1995. @
motifs common to virtually all other members of the GTPase superfamily (Bourne et al., 1991; Burns et al., 1993) and which crystallographic studies have confirmed are involved in nucleotide binding (Jurnak, 1985; Pai et al., 1990; Noel et al., 1993). Particular interest has focused on the ,6-tubulin GTPase because of its proposed role in microtubule dynamic properties and in the cellular functions of microtubules (Mitchison & Kirschner, 1984; Kirschner & Mitchison, 1986). Since it has not been possible to crystallize the tubulins, attempts to identify P-tubulin sequences involved in GTP binding have been limited primarily to nucleotide cross-linking studies and to sequence comparisons with GTPase superfamily members. Unfortunately, these studies from several laboratories have not provided a coherent picture but instead have identified many different regions (Figure 1). Amino acids P3-19 have been identified by two groups using UV cross-linking of unmodified GTP (Shivanna et al., 1993) or 8-azido-GTP (Jayaram & Haley, 1994) to the tubulin dimer. It has also been proposed that part of this region is homologous to the consensus GXXXXGK phosphate-binding motif found in other members of the GTPase superfamily (Shivanna et al., 1993). The P-tubulin sequences in this region, however, are not conserved in all P-tubulins (Burns, 1991; Little & Seehaus, 1988), as might be expected of a region involved in such a highly conserved function. Another study, which
0006-2960/95/0434-7409$09.00/0 0 1995 American Chemical Society
7410 Biochemistry, Vol. 34, No. 22, 1995
Sage et al.
Table 1: Yeast Strains Used in This Study strain /?-tubulin genotype
selection marker genotype
FY41
TUB2
leu2dl trpld63 ura3-52 his4-917
ADYlOl ADY 103 CSY3 N298K N298Q CSY3 + pCS3-D295N ADY 101-D203S
tub2-590/tub2-590 TUBZflUB2 tub2-590/tub2-590::LEU2 tub2-N298K tub2-N298Q t~b2-D295N/t~b2-590 tub2-D203S/tub2-590
HIS4his4 leu2neu2 LYS4/lys4 trpl/trpl uraYura3 HlS4his4 leu2Aeu2 LYS4/lys4 trpl/trpl ura3/ura3 HIS4/his4 Ieu2Aeu2/LEU2 LYS4/lys4 trplltrpl ura31ura3 leu2dl trpld63 ura3-52 his4-917 URA3 leu2dl trpld63 ura3-52 his4-917 URA3 HIS4his4 leu2Aeu2/LEU2 LYS4/lys4 trpl/trpl ura3/ura3/URA3 HlS4his4 leu2Aeu2 LYS4/lys4 trpl/trpl ura3/ura3/URA3
also used direct cross-linking of GTP to the tubulin dimer, identified the region consisting of amino acids ,663-77 (Linse & Mandelkow, 1988); this region was also identified using a photoaffinity analogue of GTP (Chavan et al., 1990). It has been proposed that this region is similar to a nucleotide-binding domain of EF-Tu (Linse & Mandelkow, 1988), although the homology is weak. Yet a third region, consisting of amino acids Pl55- 174, was also identified in a GTP-cross-linking study (Hesse et al., 1987). This region shows no direct homology with other nucleotide-binding proteins but is located just C-terminal to a region of P-tubulin that is very similar to other nucleotide-binding proteins (Baker et al., 1992). A second approach to identifying P-tubulin GTP-binding domains has involved comparisons of P-tubulin primary sequences with those of other nucleotide-binding proteins. By comparison with EF-Tu, Linse and Mandelkow (1988) identified the region 60KYVPRAILVD69of P-tubulin as being involved with guanine base binding; however, this region is not fully conserved in P-tubulins (Burns, 1991). The regions ‘40GGGTGSG’46 and 178TVVE’80, which are highly conserved in all tubulins, have also been identified as being involved with phosphate and ribose binding, respectively (Mandelkow et al., 1988). Sternlicht et al. (1987) proposed that the tubulins may resemble other GTPase superfamily members if the unusual step were taken of reversing the orientation of the superfamily nucleotide-bindingconsensus motifs. Thus, the region Io3KGHYTEGio9was proposed to be homologous to the superfamily phosphate-binding motif, GXXXXGK, except that its orientation was reversed in P-tubulin. Similarly, the P-tubulin region 295DAKN298 was proposed to be equivalent to the base-binding consensus motif, NKXD, of superfamily members and P-tubulin 203DNEA206 to be equivalent to the phosphate-binding motif DXXG/A. By homology with ATPases these authors also identified @-tubulin I4OGGGTGSG’46as a potential phosphate-binding domain. Interestingly, the regions identified by this approach are very highly conserved in all P-tubulins (Bums, 1991; Burns et al., 1993), and mutations site-directed to the yeast ‘03KGHYTEG’09 peptide affected both the assembly-dependent hydrolysis of GTP and microtubule dynamic properties in vitro (Davis et al., 1994). In the present study, we have extended this site-directed mutagenesis of the yeast singlecopy P-tubulin gene, TUB2, to test whether point mutations in the 295DAKN298 region affect the affinity and specificity of @-tubulinfor GTP binding. We also examined whether a D203S mutation in a proposed phosphate-binding motif of P-tubulin (203DNEA206) affected GTP-binding affinity, the microtubule assembly-dependent GTPase activity, and the dynamic properties of microtubules in vitro. The data also
reference F. Winston, personal communication Sage et al., 1995 Sage et al., 1995 Sage et al., 1995 this study this study this study this study
provide a comparison to the results of a mutagenesis study by Farr and Sternlicht (1992), who used an unusual method to assay the effects of mutations in these sequences.
MATERIALS AND METHODS Yeast and Bacterial Strains. The yeast strains used in this study are listed in Table 1. Construction of the mutant strains is described below. The growth conditions of the yeast strains, as well as those of Escherichia coli, have been described previously (Sage et al., 1995). Site-Directed Mutagenesis of TUB2. All mutations were created by the method of Kunkel et al. (1987) using mutagenic oligonucleotides on copies of TUB2 in the vector pCS3 (Sage et al., 1995). The mutated copies of TUB2 were introduced into yeast strains carrying the tub2-590 version of the P-tubulin gene, in which the coding sequence for the C-terminal 12 amino acids had been deleted (Katz & Solomon, 1988). Expression of full-length copies of the P-tubulin gene (mutated or wild type as appropriate) in the tub2-590 background and the incorporation of the P-tubulins into microtubules were followed immunologically with a rabbit polyclonal antibody specific for the C-terminal 11 amino acids (termed “anti-tail”). For mutations tub2-D203S, tub2-N298K, and tub2-N298Q, fragment-mediated gene replacement of genomic TUB2 was carried out using 4.9 kb SacI-SphI fragments of mutagenized pCS3. The fragments were transformed into the homozygous tub2-590/tub2-590 diploid strain ADYlOl (Table 1) by the method of Ito et al. (1983). Transformants were selected on the basis of conversion to uracil prototrophy and were screened for the presence of the mutant sequence by polymerase chain reaction (PCR) sequencing of genomic DNA isolated from the transformants. For unknown reasons, the mutation tub2-D295N could not be obtained by fragmentmediated transplacement. Instead, tub2-D295N was introduced into a hemizygote strain derived from ADYlOl (CSY3, tub2-590/null, Table 1) on pCS3. This plasmid carries a CENZZZ sequence which maintains the plasmid episomally at a copy number of approximately one per cell (Fitzgerald-Hayes et al., 1982; Sage et al., 1995) and thus effectively regenerated a heterozygous condition in CSY3. Transformants were again selected on the basis of uracil prototrophy, and the presence of the tub2-D295N mutation was confirmed by sequencing plasmids “rescued” from candidate yeast strains (Sage et al., 1995). To determine whether the tubulin mutations introduced into ADYlOl by gene replacement were haploid viable, the heterozygous diploids were sporulated (Kassir & Simchen, 1991) and the tetrads dissected using a Lawrence Precision micromanipulator on an Olympus BHC phase contrast microscope. Tetrads from the tub2-N298K- and tub2-
Site-Directed Mutagenesis of Yeast ,&Tubulin N298Q-containingheterozygotesproduced four viable spores, and haploid strains carrying only the mutated copy of TUB2 were first identified on the basis of uracil prototrophy. The presence of the mutations in these haploid strains was confirmed by PCR sequencing of genomic DNA. In tetrads from tub2-D203S-containing heterozygotes only two viable spores were produced, and none carried the URA3 marker cloned into the TUB2 downstream-flanking sequences. This indicated that the tub2-D203S mutation was lethal in haploids, and analysis of this mutation was carried out in the tub2-D203Sltub2-590 heterozygote. Isolation of Yeast Tubulin. Tubulin was isolated from 60 L fermenter (Fermicell 130) cultures of yeast strains by the method of Davis et al. (1993). The purity of the isolated tubulins was determined by scanning Coomassie blue stained SDS-PAGE' gels of tubulin preparations using a LKB Ultroscan densitometer. The final purities of the tubulin preparations were approximately 80-95%, with the exception of the D295N tubulin which could only be purified to ca. 35%. In the cases of the D203S and D295N mutations, which we obtained only as heterozygous tub2-D203Sltub2-590 and tub2-D295Nltub2-590 diploids, the tubulin preparations were mixtures containing both full-length (tub2-D203S, tub2D295N) and truncated (tub2-590)P-subunits. Quantitative analysis of Coomassie blue stained SDS-PAGE gels showed that the full-length mutant P-tubulins were present at levels similar to full-length wild-type P-tubulin in a TUBZltub2590 control strain (data not shown). Determination of Nucleotide-Binding Afinity . GTP-binding assays for yeast tubulins were performed by the method of Hummel and Dryer (1962) as modified for tubulin by Levi et al. (1974), using 12 cm x 0.5 cm columns of P-10 (BioRad, Richmond, CA). The columns were equilibrated in 0.1 M Pipes, pH 6.8,2 mM EGTA, 1 mM MgS04 (PEM), 10% (vlv) glycerol, and 5-10 000 nM r3H]GTP (110 Ci mol-', NEN). Tubulin samples (40-80 pg) in 500
6 6
0.4 zk 0.1 0.9 rt 0.2
46 & 3 90 iz 66 2 9 f 16
'500
43 f 6
"500
+
tubulin source controls bovine brain PC TUB2/TUB2 TUB2/tub2-5906
?95DAKN298 t~b2-D295N/tub2-590 tub2-N298K t~b2-N298Q
KO
+
+
v
1.0
'500
5
* 0.1
0.5 i 0.1 0.8 f 0.4 0.4 & 0.1
favor XTP binding reduce MgGTP binding reduce MgGTP binding
* 0.2
reduce MgGTP binding
203DNEA206
tub2-D203S/tub2-590
predicted effect based on p2 1"s
0.8
a Equilibrium dissociation constants for nucleotide binding were determined using the method of Hummel and Dreyer (1962) and were obtained from the abscissa intercepts of double-reciprocal plots of the data, using linear regression analysis. The maximum stoichiometries of GTP binding (v, moles of GTP bound per mole of tubulin dimer) were obtained from the ordinate intercepts of the double-reciprocal plots. Each value is the average of at least three experiments, and errors represents f l standard error of the mean. The predicted effects of mutations on tubulin-MgGTP binding are based on the results of mutagenesis studies with GTPase superfamily members, which are described in the text. From Davis et al. (1 993).
;iF
10
-
0
0.05
0
0.1
0
0.15
0.05
0.1
lz3 10
0
0
30
0.1
0.15
0.05
0.1
0.15
304f
e
I 2ol
zol 0
0.05
0.05
0.1
0.15
0
l/[GTPI (nM-*)
FIGURE 2: GTP binding to tubulins isolated from P-tubulin mutants. The different tubulin preparations (40-80 p g in
