Heterologous expression and folding analysis of a b-tubulin isotype from the Antarctic ciliate Euplotes focardii Sandra Pucciarelli 1,2 , Cristina Miceli 1 and Ronald Melki 2 1 Dipartimento di Biologia Molecolare, Cellulare e Animale, Universita ` di Camerino, Camerino, Italy; 2 Laboratoire d’Enzymologie et Biochimie Structurales, Centre National de la Recherche Scientifique, Gif-sur-Yvette, France Mammalian tubulins and actins attain their native con- formation following interactions with CCT (the cytosolic chaperonin containing t-complex polypeptide 1). To study the b-tubulin folding in lower eukaryotes, an isotype of b-tubulin (b-T1) from the Antarctic ciliate Euplotes focardii, was expressed in Escherichia coli. Folding analysis was performed by incubation of the 35 S-labeled, denatured b-T1 in the presence, or absence, of purified rabbit CCT and cofactor A, a polypeptide that stabilizes folded monomeric b-tubulin. We show for the first time in protozoa that b-tubulin folding is assisted by CCT and requires cofactor A. In addition, we observed that E. focardii b-T1 competes with human b5 tubulin isotype for binding to CCT. The affinity of CCT to E. focardii b-T1 and b5 tubulin are compared. Finally, the mito- chondrial chaperonin mt-cpn60 binds to b-T1 but is unable to release it in a native or quasi-native state. Keywords: protein folding; chaperonin; CCT; cpn60; protozoa. Tubulins are highly conserved proteins of eukaryotic cells involved in many essential cellular processes including intracellular transport, cell division and motility. Together with actin filaments, microtubules are the major components of the cytoskeletal lattice [1]. Microtubules are cylindrical polymers assembled from a-andb-tubulin heterodimers. In pluricellular organisms, tubulin primary sequences are encoded by a multigenic family whose products are heterogeneous proteins that can be classified as different isotypes [2]. Compelling evidence indicated that the different tubulin isotypes affect the structure and function of micro- tubules as well as their dynamic properties [3]. Tubulin heterodimers of psychrophilic organisms are able to assemble into microtubules at temperatures below 4 °C [4], while non cold-adapted microtubules disassemble. Compared with those of homeothermic animals, tubulin sequences from evolutionary distant psychrophilic organ- isms (as protozoa and fishes) revealed the presence of unique substitutions, some of which are common to all a-or b-tubulin chains with others variably distributed in the different isotypes. These substitutions, with particular post- translational modifications, may be responsible of the peculiar dynamic properties of microtubules from cold- living organisms [4–6, S. Pucciarelli and C. Miceli, unpub- lished results]. The efficient biogenesis of tubulins, as well as of actins, depends on the eukaryotic chaperonin referred to as CCT (cytosolic chaperonin containing TCP-1), TCP-1 complex, TRiC or Ct-cpn60 [7,8]. Similar to its prokaryotic counter- part GroEL, CCT has a double ring structure. However CCT has an eightfold symmetry and is composed of seven to nine distinct polypeptides (designated as a, b, c, etc.) [8] whereas GroEL has a sevenfold symmetry and is made of a single polypeptide chain. In addition, and in contrast to GroEL, CCT binds and folds a small range of misfolded polypeptides, most of which are components of the eukaryotic cytoskeleton. Misfolded actin and tubulin appear to bind to the apical domain of the CCT ring [9,10] in a nearly native conformation [10,11]. Substrate binding to CCT is accompanied by a change in the conformation of CCT [9,10,12]. Nucleotide exchange and hydrolysis also induce a change in the conformation of CCT, which increases the affinity for misfolded target proteins (as reviewed in Melki [13]). Once released from CCT, a-andb-tubulin chains interact with additional cofactors (five were discovered in mammals, denoted A–E) to constitute the native tubulin heterodimer, i.e. the functional component of microtubules [14–16]. Cofactor A (Cof A) has been shown to stabilize the b-tubulin polypeptide chain [15]. Homologues of the mammalian postchaperonin tubulin folding cofactors have been identi- fied in two species of yeast, and their function is correlated with the biogenesis of a-andb-tubulin and the constitution of functional microtubules [17,18]. In protozoa, CCT has been characterized at the gene sequence level in the ciliate Tetrahymena pyriformis [19,20], in the diplomonadide Giardia lamblia and the parabasilide Trichomonas vaginalis [21]. From these analyses the genes encoding CCT subunits in protozoa appeared homologous to those known in higher organisms. The evidence that even in early divergent eukaryotes CCT is encoded by a multigenic family suggests an ancient paralogy within eukaryotes. The predilection of CCT for cytoskeletal Correspondence to Sandra Pucciarelli, Dipartimento di Biologia M.C.A., Universita ` di Camerino, Via Camerini 2, 62032 Camerino (MC), Italy. Fax: + 39 0737 636216, Tel.: + 39 0737 413276, E-mail: sandra.pucciarelli@unicam.it Abbreviations: CCT, cytosolic chaperonin containing TCP-1; Cof A, cofactor A; mt-cpn60, mitochondrial chaperonin 60 (the number referring to the approximate kilodalton molecular mass of the sub- units); TCP-1, t-complex polypeptide 1; TriC, TCP-1 ring complex. (Received 10 September 2002, revised 28 October 2002, accepted 4 November 2002) Eur. J. Biochem. 269, 6271–6277 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03346.x proteins has led to the hypothesis of a possible coevolution between CCT and actins and tubulins [22]. Euplotes focardii is a ciliated protozoan endemic to Antarctic coastal seawater, which shows optimal survival and multiplication rates at 4–5 °C [23]. In contrast to other unicellular eukaryotes, in which the tubulin pool is generally represented by a single isotype, in E. focardii, three distinct isotypes of b-tubulin have been identified, named b-T1, b-T3 and b-T4 [15,24,25]. b-T1 appears to be the most conserved as it shows the highest degree of amino acid identity (96%) with the Euplotes b-tubulin consensus sequence, identified by the alignment of tubulin sequences of non cold-adapted congeneric species available in the GenBank database. By contrast, b-T3 and b-T4 appeared quite divergent, as the percentage of identity of both isotypes compared to the Euplotes consensus sequence is 86% (S. Pucciarelli and C. Miceli, unpublished results.). Thesynthesisofthreedifferentb-tubulin isotypes may be an adaptive strategy of E. focardii to generate an hetero- geneous pool of molecules, each one with peculiar properties, to allow microtubule polymerization at the stringent temperature conditions of the Antarctic environ- ment [25,26, S. Pucciarelli and C. Miceli, unpublished results]. Whether tubulin folding assisted by CCT plays a role in microtubular cold-adaptation has not been inves- tigated so far. Here we present the first in vitro characterization of the folding mechanism of E. focardii b-tubulin. In this study we use the most conserved isotype, b-T1, with the long-term objective to compare the folding of all the E. focardii b-tubulin isotypes with those of non cold-adapted organ- isms. We show for the first time that b-tubulin from lower eukaryotes needs the assistance of CCT complex to attain its native tridimensional structure, in a manner similar to mammalian b-tubulin. Moreover, Cof A is required to stabilize folded b-T1. These results support the hypothesis that the CCT-mediated folding evolved early in the eukaryotic lineage. MATERIALS AND METHODS Cells Cell cultures of the E. focardii strain TN1 [23] were used. They were isolated from sediment and sea water samples collected from the coastal waters of Terra Nova Bay (temperature, ) 1.8 °C; salinity, 35%; pH, 8.1–8.2) and were grown in a cold room at 4 °C, using the green alga Dunaliella tertiolecta, or the bacterium Escherichia coli,asa nutrient source. E. focardii cytoplasmic extract preparation, SDS/PAGE and immunoblotting Cytoplasmic extracts were obtained from E. focardii cells deciliated by vigorous shaking in 4% EtOH. Cell bodies were recovered by centrifugation at 800 g,4°C, suspended in PHEM buffer (60 m M Pipes-NaOH, 25 m M Hepes, 10 m M EGTA, 2 m M MgCl 2 , pH 6.9) containing protease inhibitors (5 m M EGTA, 5 m M EDTA, 2 m M phenyl- methanesulfonyl fluoride, 2 m M o-phenanthroline, 10 mgÆmL )1 pepstatin A, 5 mgÆmL )1 leupeptin), and soni- cated twice for 5 s each (14/30 amplitude microns, Soniprep 150). Each suspension was centrifuged at 14 100 g and the supernatant used as cytoplasmic extract. SDS/PAGE was performed according to the method of Laemmli [26]. After electrophoresis, the gels (10% acryl- amide) were blotted as described previously [25]. Immuno- blotting was performed using polyclonal antibodies directed against the a subunit of hamster CCT (anti-TCP-1a [12]), peroxidase–conjugated secondary antibodies (Bio-Rad) and then detected by enhanced chemiluminescence (ECL, Amersham). Subcloning of E. focardii b-T1 into the expression vector pET 11a The complete b-T1 gene sequence is available on Gene- Bank TM with accession number S72098. Due to the deviation of ciliates from the universal genetic code [27] there is still uncertainty about the amino acid at position 21 encoded by a TAG codon, that may specify glutamine or tryptophan. Therefore, in order to subclone into the expression vector pET11a (NdeI/BamHI cut), the pUC18 vector containing the b-T1 coding region was amplified by inverted PCR with primers identical to the b-T1 sequence, except for the TAG codon that was mutated to TGG for tryptophan, as this amino acid is conserved at position 21 in most tubulins from protozoa to mammals. Then, the pUC18 containing the mutated b-T1 was used as a template in a PCR with the oligonucleotides 5¢-CGAGCTCGG TACATATGAGAG-3¢, as forward primer, and 5¢-TCGACTCTAGAGGATCCCC-3¢, as reverse primer. These primers were properly designed to add restriction sites for NdeIandBamHI just before the initiation of translation codon ATG, and after the stop codon TAA, respectively. The expression vector pET-b-T1 (pET11a containing the b-T1 coding region) was automatically sequenced on an ABI Prism sequence analyzer Model 373A (PE Applied Biosystems) using the Big Dye Terminator Methodology (PE Applied Biosystems), to verify the correct frame for b-tubulin expression, and the mutated TGG codon. Purification of CCT, mt-cpn60 and Cof A Rabbit reticulocyte lysate was prepared as described previ- ously [12]. CCT was purified from rabbit reticulocyte lysate by the chromatographic methods described by Gao et al. [7]. Fractions containing CCT that emerged from the ATP- agarose were pooled and concentrated by ultrafiltration (Centricon 30, Amicon Inc., Beverly, MA, USA). Approxi- mately 250 lL of the concentrated material was applied to a Superose 6 column (HR 10/30, Pharmacia) equilibrated in 80 m M Mes, pH 6.8, 1 m M EGTA, 1 m M MgCl 2 and 1 m M dithiothreitol. Mt-cpn60 and Cof A purifications were performed as described previously [15,28]. In vitro folding assays and competition experiments In vitro folding assays and analysis of the reaction products on nondenaturing PAGE were performed in folding buffer (80 m M Mes, 1 m M EGTA, 1 m M MgCl 2 ,1m M GTP and 1m M ATP). Labeled, denatured b-T1 or human b5 (8 mgÆmL )1 ) was diluted 100-fold in 20 lL folding buffer alone, or containing 2.5 lg of CCT (or mt-cpn60), and/or Cof A (0.6–1 mgÆmL )1 ). Folding assays with rabbit 6272 S. Pucciarelli et al.(Eur. J. Biochem. 269) Ó FEBS 2002 reticulocyte lysate on E. focardii tubulin were performed by diluting the labeled, denatured b-T1 or human b5inthe presence of rabbit reticulocyte lysate diluted twofold in folding buffer. The reactions were incubated at 30 °Cfor 90 min and the products analyzed on 4.5% nondenaturing PAGE. Following staining in Coomassie blue and destain- ing, the gels were soaked in Amplify solution (Amersham), dried and autoradiographed. For competition experiments, labeled denatured b-T1 and b5 tubulins were added to increasing amounts of unlabeled b5 tubulin. The mixtures were then diluted 100 times in folding buffer containing nucleotide-free CCT (2.5 lg). After incubation for 10 min at 30 °C to allow binary complex formation, the reaction products were analyzed on nondenaturing PAGE, as described above. All reaction products were quantified by the use of a phos- phorimager. RESULTS AND DISCUSSION CCT is in the cytoplasm of E. focardii The presence of the CCT among the cellular components of E. focardii was shown by cellular fractionation and subse- quent immunoblotting analysis using anti-(TCP-1a) poly- clonal Igs [12]. Two bands with apparent molecular weight close to that of the rabbit CCT a-subunit were recognized by the antibodies in the cytosolic fraction of E. focardii (Fig. 1). One of the polypeptides presumably corresponds to the a-subunit of E. focardii CCT because it has a molecular weight identical to that of the a-subunit of rabbit CCT; the second polypeptide may correspond to another CCT subunit of E. focardii that contains epitope(s) similar to that/those of the a-subunit. These results indicate that a chaperonin containing two vertebrate TCP-1a related polypeptide chains is present in the cytosol of E. focardii. Properties of recombinant E. focardii b-T1 A number of tubulin polypeptide chains from various exotic origins [29,30] have been successfully expressed in E. coli as soluble proteins. In contrast, recombinant tubulin polypeptide chains from various vertebrates [31–33] form inclusion bodies within E. coli presumably because they are misfolded. To determine whether E. focardii b-T1 is soluble when expressed in E. coli 35 S-labeled b-T1 was synthesized as reported by Gao et al. [31] using the construct described in the Materials and methods section and analysed by SDS/ PAGE. [ 35 S]b-T1 has an apparent molecular mass indistin- guishable from that of native b-tubulin from brain or recombinant human b5 tubulin. After centrifugation at 20 000 g for 15 min at 4 °C, 96% of the [ 35 S]b-T1 was recovered in the pellet of the bacterial lysate. We conclude from these observations that b-T1 forms inclusion bodies as do mice or human b5 tubulins. This result strongly suggests that b-T1 is misfolded upon expression in E. coli. To further document b-T1 folding, pelleted [ 35 S]b-T1 was recovered, dissolved in 7.5 M urea as described [31] to a final concentration of 5 mgÆmL )1 and diluted 100 times from denaturant in folding buffer (0.1 M Mes, pH 6.8, 1 m M EGTA, 1 m M MgCl 2 ,1m M GTP, 1 m M ATP). The reaction products were analyzed on nondenaturing PAGE as described [31], immediately after dilution or after 1 h of incubation at 30 °C. b-T1 behaviour was similar to that of b5 tubulin [33]. At early times, the bulk of the radioactivity migrates as a broad band with a slower mobility than folded b5 tubulin run under identical conditions. After 1 h, virtually all of the input radioactivity was found at the origin, i.e. at the top of the gel (data not shown). We conclude that b-T1 aggregates rapidly in solution under the conditions used in our folding assays, as does b5 tubulin [33]. We further conclude that b-T1 folding properties differ significantly from that of Giardia duodenalis and Reticu- lomyxa filosa tubulins [29,30] and are instead similar to that of higher vertebrate b-tubulin. Folding of E. focardii b-T1 requires CCT and Cof A Up to now, the folding pathway of protozoan tubulins has never been established. In Tetrahymena, coexpression of genes encoding CCT subunits and tubulin during cilia recovery has been demonstrated [34,35]. However, no direct evidence for the participation of the CCT complex and Cof A with b-tubulin folding process has been documented. As a first analysis of the protozoan tubulin folding, labeled, urea-denatured E. focardii b-T1 was diluted in rabbit reticulocyte lysate, that is known to be enriched in the tubulin folding machinery [7,12]. A parallel reaction, consisting in the dilution of labeled, urea-denatured human b5 tubulin in rabbit reticulocyte lysate, was performed as a reference. The reaction products were then analyzed in two separate lanes of a nondenaturing PAGE. Unlabeled native pig tubulin was run in a third slot to locate the position of native dimeric tubulin. The autoradiogram of the gel (Fig. 2A) revealed two intense bands in the lane containing the E. focardii b-T1 folding reaction (empty arrowheads). The upper band corresponds to the b-T1/CCT complex. This band is also observed in the b5 tubulin folding reaction (lane 2). The lower band migrates at the level of pig a/b- tubulin heterodimer, while folded b5 tubulin in complex with Cof A migrates at a lower position [15]. This indicates that folded, labeled b-T1 is either incorporated in the tubulin heterodimer [36–38] or that the folded b-T1/Cof A complex has a slight slower mobility than b5 tubulin/Cof A complex. This result suggests that labeled, denatured, recombinant E. focardii b-T1 is properly folded by the mammalian tubulin folding machinery as it forms either a stable complex with Cof A or exchanges against rabbit b-tubulin in the tubulin heterodimer. To determine whether the abnormal mobility of labeled, folded b-T1 is due to its incorporation into a mixed rabbit Fig. 1. Immunodetection of CCT a-subunit in the cytoplasm of E. fo- cardii. SDS/PAGE of an E. focardii cytoplasmic fraction (20 lg, lane 1) and of CCT purified from rabbit reticulocyte lysate (2 lgand6lg in lanes 2 and 3, respectively). Lanes 1 and 2 were Western blotted and immunostained by an anti-(TCP-1a) polyclonal Ig. Lane 3 was Coo- massie stained. Molecular size markers are indicated on the right. Ó FEBS 2002 Folding of b-tubulin from Euplotes by CCT (Eur. J. Biochem. 269) 6273 a-tubulin/b-T1 tubulin heterodimer or to a slower mobility of the b-T1/Cof A complex, labeled denatured b-T1 was folded in vitro by purified rabbit CCT in the presence or in the absence of Cof A, and the reaction products were analyzed on nondenaturing PAGE. Similar reactions where human b5 tubulin was substituted to b-T1 were run in parallel, as a control. The results are presented in Fig. 2B. In the absence of Cof A, b-T1 binds to CCT (upper arrow). No folded products are generated as described [14]. In the folding reaction containing both CCT and Cof A, an additional band is generated (middle arrow). A similar band (lower arrow) corresponding to b5/Cof A complex [15] is generated in the equivalent b5 tubulin control folding reaction. Therefore, the middle band probably corresponds to folded b-T1 in complex with Cof A. The slightly lower mobility of the b-T1/Cof A complex as compared to that of the b5/Cof A complex is probably due to the difference in the isoelectric point of b-T1 and human b5, calculated from their respective primary sequences (4.55 and 3.91, respect- ively, determined using the WINPEP software [39]). Finally, in a manner similar to that observed for b5 tubulin, no folded b-T1 tubulin is generated in the folding reaction containing Cof A but lacking CCT. We conclude from these data that CCT is required for the folding of b-T1. We further conclude that folded b-T1isstabilizedbyCofA,asisb5 tubulin. E. focardii b-T1 does not fold in the presence of the mt-cpn60 Mammalian unfolded tubulin polypeptide chains bind with high affinity to the prokaryotic homologues of CCT, GroEL and the mitochondrial chaperonin cpn60. How- ever, folding was demonstrated not to proceed any further, with tubulin chains being trapped on GroEL and cpn60 [15,40]. More recently, the a-andb-tubulin chains from the giant amoeba R. filosa were shown to remain soluble when expressed in E. coli following their interaction with the prokaryotic chaperonin GroEL/ES [29]; the resulting soluble tubulin is not necessarily native, however, as no assembly reaction was performed. To test whether b-T1, which shares about 85% sequence identity with mamma- lian b-tubulin, folds into a stable soluble state in the presence of GroEL or cpn60, denatured b-T1 was incuba- ted at 30 °C for 90 min in folding buffer containing GroEL or cpn60 in the absence or in the presence of Cof A. The reaction products were analyzed on nondena- turing PAGE. Figure 2C shows that b-T1 binds to cpn60, as does human b5 tubulin incubated under identical conditions. No additional bands are observed on the gel which indicates that no stable b-T1, neither in the absence nor in the presence of Cof A, is generated following the interaction of b-T1 with cpn60. We conclude from this observation that the folding of b-T1 either does not proceed, or that the released folded product is misfolded and unstable and therefore unable to interact with Cof A. Thus, b-T1 does not acquire a native-like conformation following its interaction with cpn60, in a manner similar to mammalian actin and tubulins [40]. b-T1 and human b5 tubulin compete for the same site on CCT The results obtained by the folding assay described above clearly show that b-T1 and CCT form a complex. To determine whether b-T1 binds to CCT in a manner similar to b5 tubulin despite the difference in the primary structure, we performed a competition experiment using rabbit CCT, labeled b-T1 and unlabeled b5, as described in Melki et al. [32]. In a control reaction, a competition experiment was performed using labeled and unlabeled b5. The reaction products were analyzed on nondenaturing PAGE and quantified using a phosphorimager. As shown in Fig. 3, the radioactive signal in the CCT/tubulin complex (arrow) decreases with increasing amount of unlabeled b-tubulin. This result suggests that the binding of b5 to CCT particles hinders that of b-T1 and vice versa (not shown). A quantitative analysis of the relative yields of the CCT/ b-tubulin complex formed revealed that CCT has a similar affinity for b-T1 and b5 tubulin, although slightly lower for b-T1 (compare the intensities at given unlabeled b5 : labeled b-T1 and unlabeled b5 : labeled b5 ratios). The slight difference in affinity is more apparent when the amount of bound labeled tubulin is plotted as a function of the unlabeled : labeled b-tubulin ratio, assuming that the higher value obtained is 100% binary complex. The comparison of the slopes of unlabeled b5 : labeled b-T1 and unlabeled b5 : labeled b5 competition experiments (Fig. 3b) reveals that b-T1 binds to CCT with a twofold lower affinity than b5 tubulin. This behavior may be the consequence of Fig. 2. Folding analysis of E. focardii b-T1. (A) In vitro folding reac- tions of labeled denatured b-T1 (lane 1) and b5 (lane 2) tubulins in rabbit reticulocyte lysate. The upper and lower open arrowheads in lane 1 indicate the location of b-T1/CCT complex, and the b-T1/Cof A complex, respectively. The upper, middle and lower solid arrowheads in lane 2 indicate the location of b-5/CCT complex, the a/b5 tubulin heterodimer and b5/Cof A complex, respectively. (B) In vitro folding reactions of labeled denatured b-T1 and b5 tubulins in the presence (+) or the absence (–) of purified CCT and Cof A. The upper, middle and lower arrows indicate the location of b-tubulins/CCT complex, b-T1/Cof A complex and b5/Cof A complex, respectively. (C) In vitro folding reactions of labeled denatured b-T1 and b5 tubulins, in the presence (+) or the absence (–) of cpn60 and Cof A. The arrow indicates the location of b-tubulin/cpn60 complex. 6274 S. Pucciarelli et al.(Eur. J. Biochem. 269) Ó FEBS 2002 differences in the affinity of CCT for b-T1 and b5 folding intermediates. Alternatively, this may be due to the differ- ences in the primary structures of b-T1 and b5 tubulin as CCT binds unfolded actins, tubulins and other cytosolic proteins with different affinities [33]. Regions of actin and tubulins that are preferentially recognized by CCT have been extensively investigated [41– 44]. In b-tubulin, the peptide that comprises amino acids 251–287 and the partially overlapping peptide spanning amino acids 263–384 define tubulin surface area that is most likely to interact with CCT [41–44]. A more recent analysis of the interaction between CCT and b-tubulin by cryoelec- tron microscopy allowed the identification of three regions from the N-terminal, and five from the C-terminal domains of b-tubulin, involved in its binding to CCT [22]. Polypep- tide P259–T372 is among these regions. It is considered the most important as it binds to CCT with the strongest affinity [22]. Based on the tubulin 3D model [45,46], this polypeptide includes strand S7, that has been demonstrated to be involved in the stabilization of the native tubulin monomer by establishing an intramolecular interaction with the N-terminus, and helix H10, that is implicated in a/b interdimeric longitudinal contacts (underlined by L, in Fig. 4) [45]. It also includes amino acid residues involved in lateral contacts between tubulin heterodimers in microtu- bule walls (underlined by M, in Fig. 4): the helix H9, the H10–S9 loop, and the S7–H9 loop, defined as the ÔM loopÕ [45,46]. Comparison of the primary structure of E. focardii b-T1 polypeptide P259–T372 with the same region of the Euplotes consensus sequence, obtained from the alignment of the known b-tubulin sequences for Euplotes species (accession numbers P20365 and Q08115, and S. Pucciarelli and C. Miceli, unpublished results) and three vertebrate tubulins (including human b5) (Fig. 4), revealed 18 amino acid positions where substitutions occur. The P268I substitution (shown in bold italics) is unique to E. focardii and can be correlated with microtubule cold-stability. It affects a residue in a proline-rich motif. Site-directed mutagenesis in this tubulin motif results in a weaker affinity of the mutant tubulin for CCT [43]. Additional substitutions in b-T1 that are specific to protozoa are indicated by asterisks in Fig. 4. The substitution A352S is located before the VCDIP motif (boxed in Fig. 4) that is involved in the interaction between b-tubulin and CCT [22]. This and other substitutions (labeled in pale gray in Fig. 4) reduce the hydrophobicity of tubulin 259–372 region and may be Fig. 4. Primary structure alignment of the putative domain that interacts with CCT from four b-tubulin isotypes. Tubulin polypeptides 259–372 from E. focardii b-T1 (accession number S72098) and Euplotes b-tubulin consensus sequence (Eupl. consensus, see text) were aligned with those from human b5, mouse b5 and chicken b3 tubulins (accession numbers P04450, P05218 and P09206, respectively), and mapped on tubulin secondary structure as determined by Incla ` n and Nogales [47]. Arrows and rectangles indicate b-sheets and a-helices, respectively. The labels ÔLÕ and ÔMÕ indicate residues involved in longitudinal and lateral contacts between tubulin molecules in the microtubule, respectively [47]. Amino acid substitutions in b-T1 relative to vertebrate b-tubulins are marked by asterisks and the substitution P268I, unique to E. focardii tubulin, is shown in bold italic. Apolar amino acids are shown in grey. The intensity of the grey color corresponds to the degree of amino acid side chain hydrophobicity resulting from the comparison of the residues at a specific position (paler gray ¼ less hydrophobic). Boxed areas delineate the CCT-binding motifs of tubulin characterized by Llorca et al. [22]. Fig. 3. Competition experiments between b-T1 and b5 tubulins for bindingtoCCT.(A) Labeled denatured b-T1 tubulin (b-T1*) or b5 tubulin (b5*) were mixed with increasing concentrations of unlabeled b5 and the mixture incubated with CCT. The reaction products were analyzed by nondenaturing PAGE; the ratios of labeled to unlabeled tubulins are indicated on the top of each lane and the arrow indicates the position of the b-tubulin/CCT complex. (B) The amounts of the tubulin bound to CCT were quantitated by the use of a phosphori- mager and expressed as a fraction of the maximum amount of bound labeled tubulin and plotted as a function of the ratio of labeled : un- labeled tubulin. Ó FEBS 2002 Folding of b-tubulin from Euplotes by CCT (Eur. J. Biochem. 269) 6275 responsible for the weaker binding of b-T1 to CCT, given that CCT binds target proteins in their quasi-native conformation by interaction with exposed hydrophobic amino acid residues [42]. As mentioned in the introduction, two additional b-tub- ulin isotypes were identified in E. focardii. They are denoted as b-T3 and b-T4. Comparison of the primary structure of b-tubulin isotypes from E. focardii to those of non-Antarc- tic organisms, revealed numerous amino acid substitutions that probably accumulated in order to allow microtubules polymerization and stability at the low temperature of the Antarctic sea water ([6], S. Pucciarelli and C. Miceli, unpublished results). An analysis of the putative role of these amino acid substitutions in cold adaptation by mapping them on the 3D structure of tubulin (carried out in collaboration with E. Nogales, University of California, Berkeley, USA) will be presented elsewhere. The substitu- tions located in b-T1, b-T3 and b-T4 regions implicated in the binding to CCT [22,42,44] (Fig. 4) may have coevolved with CCT surface areas, allowing the interaction of Antarctic b-tubulin isotypes with CCT in the adverse energetic conditions of the Antarctic habitat. The role of these amino acid substitutions in the folding process of E. focardii b-tubulin isotypes should therefore be investi- gated further. ACKNOWLEDGEMENTS This work was supported by the Italian ÔProgramma Nazionale di Ricerca in AntartideÕ,bytheÔCentre National de la Recherche ScientifiqueÕ and by the ÔAssociation pour la Recherche sur le CancerÕ. REFERENCES 1. Hyams, J.S. & Lloyd, C.W. (1993) Microtubules. In Modern Cell Biology (Harford, J.B., ed.). Wiley-Liss, New York, USA. 2. Luduena, R.F. (1998) Multiple forms of tubulin: different gene products and covalent modifications. Int. Rev. Cytol. 178,207– 275. 3. Panda,D.,Miller,H.P.,Banerjee,A.,Luduena,R.F.&Wilson, L. (1994) Microtubule dynamics in vitro are regulated by the tubulin isotype composition. Proc. Natl Acad. Sci. USA 91, 11358–11362. 4. Detrich, H.W., Parker, S.K., Williams, R.C., Nogales, E. & Downing, K.H. (2000) Cold adaptation of microtubule assembly and dynamics: structural interpretation of primary sequence changes present in the a and b-tubulins of Antarctic fishes. J. Biol. Chem. 275, 37038–37048. 5. Miceli, C., Ballarini, P., Di Giuseppe, G., Valbonesi, A. & Luporini, P. (1994) Identification of the tubulin gene family and sequence determination of one b-tubulin gene in a cold- poikilotherm protozoan, the Antarctic ciliate Euplotes focardii. J. Euk Microbiol. 42, 430–437. 6. Pucciarelli, S. & Miceli, C. (2002) Characterization of the cold- adapted a-tubulin from the psychrophilic ciliate Euplotes focardii. Extremophiles 5, 385–389. 7. Gao, Y., Thomas, J.O., Chow, R.L., Lee, G H. & Cowan, N.J. (1992) A cytoplasmic chaperonin that catalyzes b-actin folding. Cell 69, 1044–1050. 8. Kubota, H., Hynes, G. & Willison, K. (1995) The chaperonin containing t-complex polypeptide 1 (TCP-1). Multisubunit machinery assisting in protein folding and assembly in the eukaryotic cytosol. Eur. J. Biochem. 230, 3–16. 9. Llorca, O.E., McCormack, A., Hynes, G., Grantham, J., Cordell, J., Carrascosa, J.L., Willison, K.R., Fernandez, J.J. & Valpuesta, J.M. (1999) Eukaryotic type II chaperonin CCT interacts with actin through specific subunits. Nature 412, 693–696. 10. Llorca, O., Martin-Benito, J., Ritco-Vonsovici, M., Willison, K.R., Carrascosa, J.L. & Valpuesta, J.M. (2000) Eukaryotic chaperonin CCT stabilizes actin and tubulin folding intermediates in open quasi-native conformations. EMBO J. 15, 5971–5979. 11. Llorca, O., Martin-Benito, J., Ritco-Vonsovici, M., Grantham, J., Hynes, G.M., Willison, K.R., Carrascosa, J.L. & Valpuesta, J.M. (2001) The Ôsequential allosteric ringÕ mechanism in the eukaryotic chaperonin-assisted folding of actin and tubulin. EMBO J. 20, 4165–4175. 12. Melki, R., Batelier, G., Soulie, S. & Williams, R.C. Jr (1997) Cytoplasmic chaperonin containing TCP-1: structural and func- tional characterization. Biochemistry 36, 5817–5826. 13. Melki, R. (2001) Nucleotide-dependent conformational changes of the chaperonin containing TCP-1. J. Struct. Biol. 135, 170–175. 14.Gao,Y.,Melki,R.,Walden,P.D.,Lewis,S.A.,Ampe,C., Rommelaere, H., Vandekerckhove, J. & Cowan, N.J. (1994) A novel cochaperonin that modulates the ATPase activity of cyto- plasmic chaperonin. J. Cell Biol. 125, 989–996. 15. Melki, R., Rommelaere, H., Leguy, R., Vandekerckhove, J. & Ampe, C. (1996) Cofactor A is a molecular chaperone required for b-tubulin folding: functional and structural characterization. Biochemistry 35, 10432–10445. 16. Tian, G., Bhamidipati, A., Cowan, N.J. & Lewis, S.A. (1999) Tubulin folding cofactors as GTPase-activating proteins. GTP hydrolysis and the assembly of the a/b-tubulin heterodimer. J. Biol. Chem. 274, 24154–24158. 17. Lopez-Fanarraga, M., Avila, J., Guasch, A., Coll, M. & Zabala, J.C. (2001) Review: postchaperonin tubulin folding cofactors and their role in microtubule dynamics. J. Struct. Biol. 135, 219–229. 18. Radcliffe, P.A., Hirata, D., Vardy, L. & Toda, T. (1999) Func- tional dissection and hierarchy of tubulin-folding cofactor homologues in fission yeast. Mol. Biol. Cell. 10, 2987–3001. 19. Soares, H., Cyrne, L., Casalou, C., Ehmann, B. & Rodrigues- Pousada, C. (1997) The third member of the Tetrahymena CCT subunit gene family, TpCCTa, encodes a component of the hetero-oligomeric chaperonin complex. Biochem. J. 326, 21–29. 20. Domingues, C., Soares, H., Rodrigues-Pousada, C. & Cyrne, L. (1999) Structure of Tetrahymena CCT theta gene and its expres- sion under colchicine treatment. Biochim. Biophys. Acta 1457, 453–459. 21. Archibald, J.M., Logsdon, J.M. Jr & Doolittle, W.F. (2000) Origin and evolution of eukaryotic chaperonins: phylogenetic evidence for ancient duplications in CCT genes. Mol. Biol. Evol. 17, 1466– 1476. 22. Llorca, O., Martin-Benito, J., Gomez-Puertas, P., Ritco-Vonso- vici, M., Willison, K.R., Carrascosa, J.L. & Valpuesta, J.M. (2001) Analysis of the interaction between the eukaryotic chaperonin CCT and its substrates actin and tubulin. J. Struct. Biol. 135,205– 218. 23. Valbonesi, A. & Luporini. P. (1990) A new marine species of Euplotes (Ciliophora, Hypotrichida) from Antarctica. Bull. Br. Mus. Nat. His. (Zool.) 56, 57–61. 24. Miceli, C., Pucciarelli, S., Ballarini, P., Valbonesi, A. & Luporini, P. (1996) Cold-stable microtubules of the Antarctic ciliate Euplotes focardii.InAntarctic Communities (B.Battaglia & J.Walton, eds), pp. 300–306. Cambridge University Press, London, UK. 25. Pucciarelli, S., Ballarini, P. & Miceli, C. (1997) Cold-adapted microtubules: characterization of tubulin posttranslational mod- ifications in the Antarctic ciliate Euplotes focardii. Cell Motil. Cytoskeleton 38, 329–341. 26. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 15, 680–685. 27. Lozupone, C.A., Knight, R.D. & Landweber, L.F. (2001) The molecular basis of nuclear genetic code change in ciliates. Curr. Biol. 11, 65–74. 6276 S. Pucciarelli et al.(Eur. J. Biochem. 269) Ó FEBS 2002 28. Viitanen, P.V., Lorimer, G.H., Seetharam, R., Gupta, R.S., Oppenheim, J., Thomas, J.O. & Cowan, N.J. (1992) Mammalian mitochondrial chaperonin 60 functions as a single toroidal ring. J. Biol. Chem. 267, 695–698. 29. Linder, S., Schliwa, M. & Kube-Granderath, E. (1998) Expression of Reticulomyxa filosa a-andb-tubulins in Escherichia coli yields soluble and partially correctly folded material. Gene 212, 87–94. 30. MacDonald, L.M., Armson, A., Thompson, R.C. & Reynoldson, J.A. (2001) Expression of Giardia duodenalis b-tubulin as a soluble protein in Escherichia coli. Protein Expr Purif 22, 25–30. 31. Gao, Y., Vainberg, I.E., Chow, R.L. & Cowan, N.J. (1993) Two cofactors and cytoplasmic chaperonin are required for the folding of a and b-tubulin. Mol. Cell. Biol. 13, 2488–2485. 32. Melki, R., Vainberg, I., Chow, R.L. & Cowan, N. (1993) Cha- peronin-mediated folding of vertebrate actin-related protein and c-tubulin. J. Cell Biol. 122, 1301–1310. 33. Melki, R. & Cowan, N. (1994) Facilitated folding of actins and tubulins occurs via a nucleotide–dependent interaction between cytoplasmic chaperonin and distinctive folding intermediates. Mol. Cell. Biol. 14, 2895–2904. 34. Soares, H., Penque, D., Mouta, C. & Rodrigues-Pousada, C. (1994) A Tetrahymena orthologue of the mouse chaperonin sub- unit CCTc and its coexpression with tubulin during cilia recovery. J. Biol. Chem. 269, 29299–29307. 35. Cyrne, L., Guerreiro, P., Cardoso, A.C., Rodrigues-Pousada, C. & Soares, H. (1996) The Tetrahymena chaperonin subunit CCT eta gene is coexpressed with CCTc gene during cilia biogenesis and cell sexual reproduction. FEBS Lett. 383, 277–283. 36. Yaffe, M.B., Levison, B.S., Szasz, J. & Sternlicht, H. (1988) Expression of a human a-tubulin: properties of the isolated sub- unit. Biochemistry 27, 1869–1880. 37. Zabala, J.C. & Cowan, N.J. (1992) Tubulin dimer formation via the release of a-andb-tubulin monomers from multimolecular complexes. Cell. Motil. Cytoskeleton. 23, 222–230. 38. Fontalba, A., Paciucci, R., Avila, J. & Zabala, J.C. (1993) Incorporation of tubulin subunits into dimers requires GTP hydrolysis. J. Cell. Sci. 106, 627–632. 39. Hennig, L. (1999) WinGene/WinPep: User-friendly software for the analysis of aminoacid sequences. Biotechniques 26, 1170–1172. 40. Tian, G., Vainberg, I.E., Tap, W.D., Lewis, S. & Cowan, N.J. (1995) Specificity in chaperonin-mediated protein folding. Nature 375, 250–253. 41. Dobrzynski, J.K., Sternlicht, M.L., Farr, G.W. & Sternlicht, H. (1996) Newly-synthesized b-tubulin demonstrates domain–specific interactions with the cytosolic chaperonin. Biochemistry 35, 15870–15882. 42. Rommelaere, H., De Neve, M., Melki, R., Vandekerckhove, J. & Ampe, C. (1999) The cytosolic class II chaperonin CCT recognizes delineated hydrophobic sequences in its target proteins. Biochemistry 38, 3247–3257. 43. Dobrzynski, J.K., Sternlicht, M.L., Peng, I., Farr, G.W. & Sternlicht, H. (2000) Evidence that b-tubulin induces a confor- mation change in the cytosolic chaperonin which stabilizes bind- ing: implications for the mechanism of action. Biochemistry 39, 3988–4103. 44. Ritco-Vonsovici, M. & Willison, K.R. (2000) Defining the eukaryotic cytosolic chaperonin-binding sites in human tubulins. J. Mol. Biol. 304, 81–98. 45. Lowe, J., Li, H., Downing, K.H. & Nogales, E. (2001) Refined structure of a b-tubulin at 3.5 A ˚ resolution. J. Mol Biol. 313, 1046– 1057. 46. Meurer-Grob, P., Kasparian, J. & Wade, R.H. (2001) Micro- tubule structure at improved resolution. Biochemistry 41, 8000– 8008. 47. Incla ` n, Y.F. & Nogales, E. (2001) Structural models for the self-assembly and microtubule interactions of c, d and e-tubulin. J. Cell. Sci. 114, 423–432. Ó FEBS 2002 Folding of b-tubulin from Euplotes by CCT (Eur. J. Biochem. 269) 6277 . Heterologous expression and folding analysis of a b-tubulin isotype from the Antarctic ciliate Euplotes focardii Sandra Pucciarelli 1,2 , Cristina Miceli 1 and. ciliate Tetrahymena pyriformis [19,20], in the diplomonadide Giardia lamblia and the parabasilide Trichomonas vaginalis [21]. From these analyses the genes encoding