Báo cáo khoa học: Different roles of two c-tubulin isotypes in the cytoskeleton of the Antarctic ciliate Euplotes focardii Remodelling of interaction surfaces may enhance microtubule nucleation at low temperature doc

Comparison of the c-T1 and c-T2 pri-mary sequences to a Euplotes c-tubulin consensus, derived from mesophilic i.e.. Structural models of c-T1 and c-T2, obtained using the 3D structure of

cytoskeleton of the Antarctic ciliate Euplotes focardii

Remodelling of interaction surfaces may enhance microtubule

nucleation at low temperature

Francesca Marziale1, Sandra Pucciarelli1, Patrizia Ballarini1, Ronald Melki2, Alper Uzun3,

Valentin A Ilyin3, H W Detrich III3 and Cristina Miceli1

1 Dipartimento di Biologia Molecolare, Cellulare e Animale, University of Camerino, Italy

2 Laboratoire d’Enzymologie et Biochimie Structurales, Centre National de la Recherche Scientifique, Gif-sur-Yvette, France

3 Department of Biology, Northeastern University, Boston, MA, USA

Microtubule assembly in metazoan cells is nucleated

by organizing centers, which include centrioles, basal

bodies, and other structures Mitotic centrosomes

con-tain a pair of centrioles and associated pericentriolar

material, whereas basal bodies recruit other accessory

structures [1,2] Both centrioles and basal bodies require c-tubulin, the ubiquitous third member of the

‘tubulin superfamily’ [3–5], for their assembly and maintenance [6–8], and for their capacity to nucleate microtubules [9] This tubulin variant associates with

Keywords

microtubule nucleation; molecular

cold-adaptation; psychrophilic microorganism;

quantitative PCR; tubulin genes

Correspondence

C Miceli, Dipartimento di Biologia

Molecolare, Cellulare e Animale, University

of Camerino, Via Gentile III da Varano,

62032 Camerino (MC), Italy

Fax: +39 0737 40 32 90

Tel: +39 0737 40 32 55

E-mail: cristina.miceli@unicam.it

(Received 22 July 2008, revised 27 August

2008, accepted 4 September 2008)

doi:10.1111/j.1742-4658.2008.06666.x

c-Tubulin belongs to the tubulin superfamily and plays an essential role in the nucleation of cellular microtubules In the present study, we report the characterization of c-tubulin from the psychrophilic Antarctic ciliate Eupl-otes focardii.In this organism, c-tubulin is encoded by two genes, c-T1 and c-T2, that produce distinct isotypes Comparison of the c-T1 and c-T2 pri-mary sequences to a Euplotes c-tubulin consensus, derived from mesophilic (i.e temperate) congeneric species, revealed the presence of numerous unique amino acid substitutions, particularly in c-T2 Structural models of c-T1 and c-T2, obtained using the 3D structure of human c-tubulin as a template, suggest that these substitutions are responsible for conformational and⁄ or polarity differences located: (a) in the regions involved in longitudi-nal ‘plus end’ contacts; (b) in the T3 loop that participates in binding GTP; and (c) in the M loop that forms lateral interactions Relative to c-T1, the c-T2 gene is amplified by approximately 18-fold in the macronuclear gen-ome and is very strongly transcribed Using confocal immunofluorescence microscopy, we found that the c-tubulins of E focardii associate throughout the cell cycle with basal bodies of the non-motile dorsal cilia and of all of the cirri of the ventral surface (i.e adoral membranelles, paraoral mem-brane, and frontoventral transverse, caudal and marginal cirri) By contrast, only c-T2 interacts with the centrosomes of the spindle during micronuclear mitosis We also established that the c-T1 isotype associates only with basal bodies Our results suggest that c-T1 and c-T2 perform different functions

in the organization of the microtubule cytoskeleton of this protist and are consistent with the hypothesis that c-T1 and c-T2 have evolved sequence-based structural alterations that facilitate template nucleation of microtu-bules by the c-tubulin ring complex at cold temperatures

Abbreviations

qPCR, quantitative PCR; RATE, rapid amplification of telomeric ends; TuRC, tubulin ring complex.

other proteins to form two macromolecular structures,

the c-tubulin small complex, which possesses a weak

microtubule nucleating activity [10,11], and the

c-tubu-lin ring complex (TuRC) [12], which nucleates

strongly c-TuRC resembles a lock washer and is

con-sidered to be the fundamental unit required for

micro-tubule nucleation Two models have been proposed to

explain microtubule nucleation by c-TuRC: (a) the

‘protofilament’ model, in which the c-tubulin subunits

of c-TuRC associate longitudinally with ab-tubulin

dimers [13], and (b) the ‘template’ model, in which the

c-TuRC ring mimics the end of a microtubule, and

c-tubulin interacts both longitudinally and laterally

with a-tubulin but only laterally with b-tubulin [14,15]

Microtubule assembly is entropically driven,

pre-dominantly via hydrophobic interactions, and therefore

is sensitive to environmental temperature both in vitro

and in vivo [16,17] The ab-tubulin dimers of

mam-mals, for example, form microtubules in vitro at

tem-peratures near 37C, and these polymers dissociate at

low temperature (4C) to yield tubulin dimers and

ring-shaped oligomers [18–20] Ectothermic

(cold-blooded) Antarctic fishes, by contrast, possess tubulins

that polymerize at temperatures as low as )1.8 C,

which is the freezing point of their chronically cold

marine habitat [16,17] Detrich and colleagues have

shown that thermal compensation of microtubule

assembly and dynamics in these fishes results from the

evolution of structural changes intrinsic to the a- and

b-tubulins [21–24]

The nucleation of cytoplasmic microtubules by

cen-trosomes requires productive binding reactions

between c-tubulin and the ab-tubulin dimer, but the

molecular alterations that conserve nucleation in

cold-living organisms have not been studied Data indirectly

relevant to temperature compensation of microtubule

nucleation were obtained from alanine-scanning

muta-genesis of the c-tubulins of Tetrahymena thermophila

[25] and Aspergillus nidulans [26] Substitution of

ala-nine at sites in the lateral surfaces (the H3 helix and

the M loop) of these c-tubulins causes cold-sensitivity

of cell growth and⁄ or loss of basal bodies [25,26] In

light of this evidence, we propose that the capacity of

c-tubulin to perform efficient microtubule nucleation

at cold temperatures reflects evolved molecular

altera-tions to its interaction surfaces

Psychrophilic ciliated protozoa are uniquely suited

to an investigation of this issue As single cells, ciliates

are directly exposed to environmental factors

through-out their life cycle, and modifications of the primary

sequences of many of their proteins are likely to reflect

adaptive mutations that increase the fitness of the

organism at cold temperatures In ciliates, microtubule

nucleation is promoted mainly by basal bodies, which are positioned precisely in organized rows in the somatic cell cortex and in the oral apparatus [8] The assembly and maintenance of basal bodies were both shown to require c-tubulin [7,8]

The ciliate Euplotes focardii, which is endemic to Ant-arctic coastal seawaters, shows strictly psychrophilic phenotypes, including optimal survival and multiplica-tion rates at 4–5C [27], the lack of a transcriptional response of the Hsp70 genes to thermal shock [28], and modifications in the primary structures of the a- and b-tubulin [29–31] and of the proteins that form the ribo-somal stalk [32] In the present study, we characterized the two c-tubulin isotypes, c-T1 and c-T2, of E focardii, model their 3D structures, and examined their differen-tial expression and cellular localization We suggest that novel amino acid substitutions located at the plus ends, near the GTP-binding sites, and within the M loops of the E focardii c-tubulins, preserve their microtubule-nucleating activities at cold temperatures and⁄ or confer different functions on the two isotypes

Results

Sequence analysis of E focardii c-tubulin genes Two c-tubulin genes (nanochromosomes), designated c-T1 (1623 bp; GenBank accession number EF189704) and c-T2 (1619 bp; GenBank accession number EF189705), were obtained by our rapid amplification

of telomeric ends (RATE)-PCR-based cloning strategy The existence of more than two c-tubulin genes in

E focardii was excluded by restriction analysis of macronuclear DNA Figure 1A shows that undiges-ted macronuclear DNA gave a single band of approxi-mately 1.6 kb (lane 1) when hybridized at low stringency to a probe derived from the c-T2 gene Co-digestion of macronuclear DNA by EcoRI and HindIII (lane 2) gave strongly hybridizing fragments of approximately 640, 480, and 300 bp, and weakly hybridizing bands of approximately 750 and 200, con-sistent with the lengths and restriction maps of the two c-tubulin nanochromosomes (Fig 1B) The restriction maps and relative abundances of the DNA fragments suggest that the c-T2 nanochromosome is amplified to

a greater extent than the c-T1 nanochromosome Both isotypes are expressed, as shown by the recovery of distinct c-T1 and c-T2 cDNAs of approximately 1.4 kb Furthermore, northern blot analysis of mRNA extracted from exponentially growing E focardii cells, when hybridized at low stringency to the c-T2 probe, indicated that the two mRNAs were comparable in size (1.4 kb; not shown)

The coding sequences of the E focardii c-T1 and

c-T2 nanochromosomes were interrupted by two

introns located in identical positions (Fig 1B) The first

intron included nucleotides 50–96 and the second intron

included nucleotides 210–253 in each gene Excluding

introns and stop codons, the c-T1 and c-T2 coding

regions were each 1383 bp in length and predicted

proteins of 461 amino acids The nucleotide sequence

identity between c-T1 and c-T2 was 94.6% Two

in-frame UGA codons, which are known to code for

cysteine in other Euplotes species [33,34], were present at

residue positions 109 and 185 in each of the genes

Comparative structural modelling of E focardii c-tubulins to human c-tubulin

The 3D structures of the E focardii c-tubulins were modeled comparatively with respect to human c-tubu-lin [35] The predicted structures of c-T1 and c-T2 were remarkably similar to that of the human protein (Fig S1)

Structural features of E focardii c-tubulin isotypes

Plus ends The deduced amino acid sequences of the c-T1 and c-T2 isotypes were aligned with respect to a Euplotes c-tubulin consensus sequence and mapped onto the consensus secondary structure of the tubulin mono-mer [35,36] (Fig 2) c-T1 and c-T2 were 95.4% iden-tical in amino acid sequence The main differences of the two isotypes compared to the Euplotes c-tubulin consensus were found in two regions, 390–403 and 70–95 (Fig 2), both of which are located at the plus end (Fig 3) In the former, c-T2 contained several polar-for-charged substitutions (K394S, R395N, D396N, and K403Q) with respect to the consensus (consensus residue⁄ residue position⁄ c-T2 residue) c-T1 displayed substitutions of bulky residues with respect to the consensus sequence (T391I, K394R; consensus⁄ position ⁄ c-T1), reciprocal changes of polar and charged amino acids (D396N, N400D), and one polar-for-hydrophobic alteration (I401N) Notable amino acid substitutions in the second region (70–95)

of c-T1 and c-T2 with respect to the Euplotes consen-sus included the V of c-T1 and K of c-T2 for G at position 76, A of c-T1⁄ c-T2 for the consensus P at position 81, G for S at position 84, F for Y at posi-tion 92, and S for A at posiposi-tion 94 Together, these results show that the plus-end surfaces of the two

E focardii c-tubulins have diverged considerably from those of mesophilic Euplotes species, with an overall tendency toward greater hydrophobicity By contrast, very few changes were observed in sequences that contribute to the c-tubulin minus end (Figs 2 and 3)

Isotypic substitutions

E focardii c-T1 and c-T2 differed considerably between themselves at their plus ends Major residue

Y398F, D400T, N401T, and K403Q (c-T1⁄ residue position⁄ c-T2) This suite of residue substitutions may confer unique functions upon each isotype

Fig 1 The macronucleus of Euplotes focardii contains two

differ-ent c-tubulin nanochromosomes (A) Southern blot analysis of the

two c-tubulin genes of E focardii using the c-T2 gene as probe.

Lane 1, undigested DNA; lane 2, EcoRI- and HindIII-digested DNA.

The sizes (bp) of DNA standards are indicated on the left The sizes

of the two c-tubulin nanochromosomes (1600 bp) and their

diges-tion products are indicated on the right (B) Structural features and

EcoRI ⁄ HindIII restriction maps of the E focardii c-T1 and c-T2

nano-chromosomes Coding, noncoding regions, introns, and telomeres

(C 4 A 4 ⁄ G 4 T 4 ) are indicated in the key.

Nucleotide-binding sites

Human c-tubulin binds GTP in a plus-end cleft

enclosed by residues G11, Q12, C13, Q16, G101,

N102, S140, A142, G143, G144, T145, V171, P173, N207, F225, I228, and N229 (where the residues shown underlined form main- and⁄ or side-chain hydrogen bonds with atoms of the nucleotide) [35]

Fig 2 Sequence comparisons of Euplotes focardii c-T1 and c-T2 with the Euplotes c-tubulin consensus The unique substitutions of E focar-dii c-T1 and c-T2 are shown as a single-letter code underneath the Euplotes c-tubulin consensus sequence; conserved residues are indicated

by dots Predicted secondary structural elements, H for helices and S for strands [37], are represented by white cylinders and black arrows, respectively T1 to T7 indicate loops that are involved in contacts with the bound GTP [37] Residues involved in longitudinal contacts at the

‘plus’ and ‘minus’ ends are indicated by ‘+’ and ‘ )’, respectively, whereas those involved in the lateral contacts of the H3 and M-loop are shown by ‘H’ and ‘M’ Regions thought to participate in binding to ab tubulin heterodimers [68] are underlined.

These residues are all conserved in the E focardii

c-tubulins Near the entrance to the nucleotide pocket

within the H2 helix, c-T2 possessed a striking

sub-stitution at position 72: glycine in place of the

c-T1⁄ consensus arginine (Figs 2–4) The presence of

glycine at this position in c-T2 has only been

observed in E focardii and in the psychrotolerant

Euplotes crassus [34, present study] By contrast,

sub-stitution of alanine for arginine 72 in the c-tubulins

of T thermophila and A nidulans produces a lethal phenotype [25,26], which suggests that this basic resi-due is important for c-tubulin function at moderate temperature The sequences of the nucleotide-binding T3 and T5 loops of c-T1 and c-T2 were highly con-served across all Euplotes species; the former perfectly, whereas the latter contained a single polar-for-hydrophobic change, I174N (consensus⁄ position ⁄ c-T1 and c-T2)

Fig 3 3D mapping of the sequence substitutions of Euplotes focardii c-T1 and c-T2 with respect to the Euplotes c-tubulin consensus (A, B) c-T1 viewed from the side and from the plus end, respectively (C, D) c-T2 viewed from the side and from the plus end, respectively Ribbon diagrams of c-T1 and c-T2 were obtained by comparative modelling to human c-tubulin using MODELLER , version 9.1 (http://www salilab.org/modeller/) [64] Residues that distinguish the E focardii c-tubulins from the Euplotes consensus are shown in red and annotated

as consensus residue ⁄ position ⁄ c-T1 or c-T2 residue The GTP molecule is shown in green The plus and minus ends, the H3 helix, and the M-loop are indicated.

M loops

The ‘extended’ M loop, which we define as

encom-passing the S7-H9 (M) loop, H9, and the H9-S8

loop, and the H3 surfaces of c-tubulin are involved

in lateral contacts [35–37] Amino acid substitutions

with respect to the Euplotes consensus were found in

the extended M loop in E focardii c-T1 and c-T2

(Figs 2 and 3) Two hydrophobic-for-hydrophobic

changes occurred near position 280 (F279L, V282I,

consensus⁄ position ⁄ c-T1 and c-T2) and c-T1

pos-sessed an alanine at 280 in place of consensus

threonine The changes in the H9-H9¢ loop were

more dramatic Both c-T2 and c-T1 contained

pro-line-for-hydroxyl substitutions (T297P and T303P,

respectively)

Tertiary structural differences between E focardii c-T1 and c-T2

Figure 4 shows the superimposition of the 3D struc-tures of c-T1 and c-T2 from the side and the plus end, respectively The comparison demonstrates that the differences between c-T1 and c-T2 (c-T1⁄ residue posi-tion⁄ c-T2) mapped largely to exposed areas (plus-end loops and helices, extended M loop) of the polypep-tides The valine at S3 position 93 of c-T2 appears to confer a conformational change in the adjacent T3 loop (Fig 4B, double arrow), which is directly involved in the formation of the GTP-binding site The alanine at

280 in c-T1 apparently causes a conformational change

in the M loop (S7-H9), which may influence lateral interactions (Fig 4A, double arrow) The substitutions

Fig 4 Comparison of the tertiary structures of Euplotes focardii c-T1 and c-T2 Ribbon diagrams of the two proteins, obtained by compara-tive modelling to human c-tubulin using MODELLER , version 9.1 (http://www.salilab.org/modeller/) [64], are superimposed to highlight structural differences The c-T1 and c-T2 loops are shown in yellow and cyan, respectively Residue substitutions that differentiate the two E focardii c-tubulins are colored violet and designated as c-T1 ⁄ residue position ⁄ c-T2 Notable loop displacements are indicated by double arrows The GTP molecule is shown in green The Mg 2+ ion (blue sphere) is shown bound to the b- and c-phosphates of GTP (dark green) (A) Side view (A¢, A¢¢) Show close-up side views (generated using Chimera; http://www.cgl.ucsf.edu/chimera/) [69] of the H9-H9¢ loop (dashed box in A), which contains a proline at position 303 in c-T1 (A¢) in contrast to the serine of c-T2 (A¢¢) The proline substitution of c-T1 eliminates the hydrogen bond between Ser303 and Asn205 in c-T2 (B) Plus-end view (B¢) An enlargement of the nucleotide-binding pocket, shown boxed in (B) Near the entrance to the pocket, c-T2 contains glycine (cyan) in place of the arginine (yellow) normally found at position 72 in helix H2.

of prolines for threonine at position 297 of c-T2 and for

serine at 303 in c-T1 do not alter significantly the

con-formation of the H9-S8 loop (Fig 4), although they are

likely to restrict its mobility However, the Pro303

sub-stitution of c-T1 eliminates the bent hydrogen bond that

forms between Ser303 and Asn205 in c-T2 (compare

Fig 4A¢,A¢¢) The cluster of substitutions in H11 and

the H11-H12 loop of the two c-tubulins cause polarity

changes at the plus end (Fig 4) that would differentiate

the longitudinal interactions formed by c-T1 and c-T2

Finally, c-T2 possesses a glycine at position 72 in place

of consensus⁄ c-T1 arginine (Fig 4B¢) This substitution

may ‘open’ the nucleotide-binding site to facilitate

exchange

We have not attempted to quantify the loop

dis-placements because the T3 and M loops of the 3.0 A˚

crystal structure of GTP-bound tubulin are disordered

[35] Hence, we consider the modeled loop

displace-ments of the E focardii c-tubulins to be provisional

and to require future validation

Transcription of the E focardii c-T1 and c-T2

nanochromosomes

To gain insight into the roles of the E focardii

c-tubu-lin isotypes, we measured the steady-state levels of

macronuclear mRNAs transcribed from the c-T1 and

c-T2 nanochromosomes of starvation-synchronized

cultures by quantitative PCR (qPCR) During

starva-tion, the transcript levels for both isotypes were low

(Fig 5A) After feeding, the amounts of c-T1 and

c-T2 mRNAs increased, with the latter being two- to

three-fold higher than the former at 18 h At 36 h

post-feeding, c-T2 mRNA increased 16-fold relative to its abundance at 18 h (53-fold increase with respect

to t = 0 h), whereas the level of the c-T1 transcript remained unchanged Ninety-eight percent of the cells were undergoing mitosis⁄ cytokinesis at this time (as determined by counting of cells using a stereomicro-scope) By 54 h, c-T2 mRNA returned to a value simi-lar to that at 18 h, whereas the amount of the c-T1 transcript was comparable to that observed at 18 and

36 h Thus, the amount of the c-T2 transcript varies widely during the cell cycle, whereas the c-T1 mRNA

is expressed at low, almost constant levels

The disparity between c-T1 and c-T2 transcript lev-els could result from differential amplification of the corresponding macronuclear nanochromosomes, as has been reported for other Euplotes genes [38], from dif-ferent rates of transcription initiation and elongation between the two genes, and⁄ or from variation in the rates of degradation of the two messages To test the first hypothesis, the gene copy number of c-T1 and c-T2 was estimated by qPCR c-T1 and c-T2 nano-chromosomes were present at approximately 175 and

3600 copies per cell, respectively (Fig 5B) Thus, the c-T2 template was approximately 21-fold more abun-dant than the c-T1 template To evaluate the second hypothesis, the 5¢ and 3¢ noncoding sequences of the two c-tubulin genes were compared Figure 5C shows that the 5¢-UTR of c-T1 contained the sequence TGA-TAC ()26 to )21; gray shading), which matches the consensus sequence for GATA-binding transcription factors (WGATAR), whereas the c-T2 5¢-UTR pos-sessed two tandem repeats ()32 to )27, )24 to )19;

gray shading) of the same motif in essentially the same

A B

C

Fig 5 Macronuclear amplification and

tran-scription of Euplotes focardii c-T1 and c-T2

genes (A) Cell-cycle-dependence of

steady-state c-T1 and c-T2 mRNA levels

deter-mined by qPCR Values are the mean ± SD

(n = 4) (B) Determination of macronuclear

gene copy-number of c-T1 and c-T2 by

qPCR Values are the mean ± SD (n = 4).

(C) Sequences of the 5¢- and 3¢ noncoding

regions of c-T1 and c-T2 putative GATA

tran-scription factor-binding motifs are shown in

gray.

location GATA-binding transcription factors are

known to regulate the transcription of some genes in

protists [39] The 3¢-UTR of c-T2 was three

nucleo-tides shorter than that of c-T1 but, otherwise, these

two sequences were quite similar We have not yet

investigated the role of message degradation with

respect to the control of c-tubulin transcript

abun-dance With the latter caveat, we propose that the

quantities of c-T1 and c-T2 mRNAs are regulated, at

least in part, by differential gene amplification and by

the number of GATA-factor promoter motifs

Distribution of c-tubulins in E focardii cells

To examine the cellular distribution of c-tubulin, we

used polyclonal anti-(human c-tubulin) serum [40] and

a polyclonal antibody that we prepared against the

most divergent peptide [(390)RIFRRRNAYIDNYK

(403)] of E focardii c-T1 Figure 6 presents confocal

microscopic images of three E focardii cells after

stain-ing with antibodies directed against a- and c-tubulins

(cell 1, Fig 6A–H; cell 2, Fig 6I–L; cell 3, Fig 6M–P)

The anti-(human c-tubulin) serum stained all classes of

basal bodies (Fig 6B,F, red), including those of: (a) the

adoral membranelles that nucleate the microtubules of

the cytostomal ciliature (labeled green by DM1A in

Fig 6A,C); (b) the paraoral membrane that surrounds

the cytostomal area; (c) the four groups of locomotory

cirri [frontoventral (numbered 1–10), transverse,

cau-dal, and marginal]; and (d) the nonmotile cilia of

the dorsal surface [41], which are arranged in

longitudi-nal rows (kineties) Interestingly, dorsal ciliary

micro-tubules were absent in the equatorial area (Fig 6E,G),

which suggests that this cell is entering mitosis and that

duplication of basal bodies at the dorsal surface

requires the disassembly of dorsal cilia

In mitotic E focardii cells (Fig 6I–P), the

anti-(human c-tubulin) serum stained newly-formed basal

bodies (Fig 6J,L, red, arrows) and the poles of the

micronuclear mitotic spindle (Fig 6J, solid

arrow-head), but macronuclear staining was never observed The basal bodies indicated by the upper arrow in Fig 6J will form the transverse, caudal and marginal cirri of the anterior daughter cell, which also inherits the frontoventral cirri of the parental cell Conversely, the basal bodies marked by the lower arrow will pro-duce the frontoventral cirri of the posterior daughter cell [42,43] and its transverse, caudal, and marginal cirri derive from the parent As division proceeds, the duplicated basal bodies nucleate new ciliary micro-tubules of the nascent cirri, as shown by the DMIA staining in Fig 6I

To determine the subcellular localization of c-T1 and c-T2, we attempted to prepare rabbit polyclonal antibodies specific for the two peptides that clearly dis-tinguish c-T1 [(390)RIFRRRNAYIDNYK(403)] and c-T2 [(390)KKLRSNNAFITTYQ(403)] We obtained

an antibody specific for c-T1 The c-T2 peptide was, however, not immunogenic Fig 6M–O shows that the anti-c-T1 serum gave staining identical to that observed with anti-(human c-tubulin), with the excep-tion that the micronuclear spindle poles were not reognized Therefore, we conclude that T2, but not c-T1, participates in the assembly of the mitotic spindle

of E focardii and that both isotypes are involved in the nucleation of other microtubule structures

Finally, we examined the distribution of E focardii c-tubulins in total cell extracts and in subfractions enriched in basal bodies or in micronuclei using anti-(human c-tubulin) and anti-c-T1 sera Figure 7 shows that the human antibody recognized c-tubulins in all three samples, whereas the c-T1 antibody gave positive signals only for the total cell extracts and basal bodies These results confirm that c-T2 alone nucleates micro-tubules in the micronucleus

Discussion

In the present study, we have shown that the psychro-philic ciliate E focardii possesses two c-tubulin genes

Fig 6 Spatial distribution of c-tubulins in Euplotes focardii cells Confocal immunofluorescence microscopic images of three E focardii cells were recorded after staining with antibodies directed against a- and c-tubulins Six optical sections, separated by intervals of 1 lm, were col-lected and merged for each cell ⁄ antigen combination (A–D) Ventral view of cell 1 in late vegetative stage; (E–H) dorsal view of cell 1; and (I–L) ventral view of cell 2 in mitosis The arrowhead in (J) indicates the micronuclear mitotic spindle, and the arrows show the newly-formed basal bodies (M–P) Ventral view of cell 3 in mitosis The arrows in (M) and (O) indicate the micronuclear mitotic spindle (A–C, E–G, I–K) Cells were co-stained with mouse monoclonal anti-a-tubulin serum DM1A (Amersham) and rabbit polyclonal anti-(human c-tubulin) serum The primary antibodies were detected using Alexa Flour 488 goat anti-(mouse IgG) (green signal indicates microtubules) and Alexa Fluor 594 goat anti-(rabbit IgG) (red signal indicates c-tubulin in basal bodies) (M–O) Cell 3 was stained for microtubules with the primary antibody DM1A and for c-tubulins with rabbit polyclonal anti-(E focardii c-T1); secondary antibodies were as before Co-localization of a- and c-tubulins

is shown by the yellow signals in merged images (C, G, K, O) (D, H, L, P) Black-and-white versions of the merged images are labeled to identify cytoskeletal structures am, adoral membranelles; pm, paraoral membrane; 1–10, frontoventral cirri involved in locomotion; tc, trans-verse cirri; cc, caudal cirri; mc, marginal cirri; mb, microtubule bundles that elongate from the basal bodies of each transtrans-verse cirrus into the cytoplasm Scale bar = 10 lm.

that encode distinct isotypes, c-T1 and c-T2 The

amino acid sequences of the two isotypes have

diverged from those of mesophilic Euplotes species

primarily in two regions that are involved in protein– protein quaternary interactions: (a) the plus end, which forms longitudinal contacts, and (b) the extended

Mitotic spindle

Mitotic spindle

Kinety

Dorsal Equatorial ar

ea

Newly formed basal bodies

cirri

M loop, which participates in lateral bonding The

extensive alterations of sequence elements that form

these surfaces are likely to be adaptations that preserve

c-tubulin function at cold temperatures Moreover,

c-T1 and c-T2 differed substantially in their sequences

at these locations, consistent with the possibility that

the functions of the single c-tubulin found in most

organisms may be partitioned between the two

protis-tan isotypes Together, our results suggest strongly

that E focardii has evolved c-tubulins that are able to

nucleate microtubule structures at low temperature

while individually performing specialized subfunctions

The E focardii c-tubulin gene family – regulation

of expression

The two c-tubulin genes of E focardii appear to be a

feature characteristic of this protistan genus Two

c-tubulin genes have also been reported for

Euplotes octocarinatus [44] and for E crassus [34] In

the former case, the c-tubulin genes produce identical

proteins, whereas, in the latter, they encode two

differ-ent isotypes whose functional differences, if any, are

unknown [34]

Transcription of the E focardii c-T2 gene was

robust and cell-cycle dependent, whereas synthesis of

the c-T1 mRNA occurred at low, almost constant

lev-els The differential transcription of the two c-tubulin

genes appears to be due to the greater copy number of

the c-T2 nanochromosome in the macronucleus (i.e

20-fold larger than that of c-T1) and to the duplication

of a GATA-transcription factor binding site in the

c-T2 promoter, although other processes might also be

involved In multicellular organisms, GATA-binding

factors play critical roles in development, including

cell-fate specification, regulation of differentiation, and

control of cell proliferation and movement [45]

Recently, we have shown that activation of the heat-shock response in the ciliate T thermophila requires GATA motifs, heat-shock transcription elements, and their cognate transcription factors [39] Taken together, our results with protistan genera support the hypothe-sis that the GATA gene-regulatory system arose early

in metazoan evolution

Sequence changes in relation to the tertiary and quaternary structures of the E focardii c-tubulins – implications for cold adaptation of tubulins Although the c-tubulins of E focardii are strikingly similar in overall 3D organization to human c-tubu-lin, the former contain divergent sequence elements that are likely to affect the mobility of their domains and their interactions with partner proteins That the plus-end surfaces of c-T1 and c-T2 differ physico-chemically from each other and from that of the mesophilic Euplotes consensus is clear The extended

M loops of c-T1 and c-T2 are more hydrophobic than the Euplotes consensus and contain proline sub-stitutions whose role may be to constrain the lateral contact residues to a conformation that is favorable for formation of the c-TuRC nucleation complex from multiple c-tubulin small complexes [12] Similarly, the lateral surfaces of a-tubulins from

E focardii and from two psychrophilic algae of the genus Chloromonas contain hydrophobic substitutions with respect to the corresponding a-isotypes of tem-perate congeners [31,46] Detrich et al [27] have shown that a small number of hydrophobic substitu-tions in Antarctic fish tubulins appear to be impor-tant for compensatory adaptation of microtubule assembly at cold body temperatures; one such change, F200Y (Antarctic fish residue⁄ position ⁄ meso-philic residue), which is located at the interface between the nucleotide-binding and intermediate domains of b-tubulin, clearly affects microtubule dynamics when mutated in Schizosaccharomyces pombe [47] Thus, increased hydrophobicity of tubu-lins, both at surface interaction sites and at internal domain interfaces, emerges as a common theme for psychrophilic organisms

Sequence alterations near the nucleotide-binding pocket of E focardii c-tubulins are also candidates for adaptive compensation Amino acid changes adjacent to the T3 loop and within the T5 loop may influence the binding affinity and hydrolysis of GTP and⁄ or the egress of GDP and Pi, which may in turn control the assembly and stability of new basal bodies [25] The most striking of these is the G72R substitution of c-T2 near the entrance to the

Fig 7 Distribution of c-tubulins in Euplotes focardii nuclei, basal

bodies, and total cell extracts (A) Total cell extracts (TCE) and

sub-fractions enriched in basal bodies (BB) and in micronuclei (N) were

prepared as described in the Experimental procedures Western

blots of the extracts and fractions were incubated with (A)

poly-clonal antibodies against human c-tubulin and (B) polypoly-clonal

antibod-ies specific for E focardii c-T1.