THE JOURNAL OF B I O L ~ ~ ~ I ~ A L CHEMISTRY 0 1986 by The American Society of Biological Chemists, Inc.

Vol. 261, No. 24, Issue of October 5, pp. 13317-13322,1986 Printed in U.S.A.

Sequence and Expression of the Chicken p3 Tubulin Gene A VERTEBRATE TESTIS B-TUBULIN ISOTUPE*

(Received for publication, May 16, 1986)

Kevin F. Sullivan$$, Paula S. MachlinS, Harry Ratrie 1119, and Don W. Clevefand$I From the D e ~ r ~ ~ n t s of ~ ~ ~ o ~ g ~ a l Chemistry and flCeU 3 ~ ~ ~ , The Johns ~ o ~ ~ i ~ ~ n ~ u e T s i ~ y School of ~ e d ~ i ~ , att ti more, ~ a r y t a n d 21205

We report the determination of the complete DNA sequence for cS3, a chicken &tubulin gene which we show to be the dominant @-tubulin expressed in testis. Like all previously studied vertebrate &tubulin genes, the gene is divided into four exon sequences inter- rupted by three intervening sequences (located be- tween amino acids 19 and 20, within codon 66, and within codon 93). Analysis of the program of expres- sion of this gene indicates that it encodes the dominant chicken testis #?-tubulin, although it is also expressed at lower levels in a wide variety of cell and tissue types. Comparison of the predicted polypeptide se- quence for cS3 with four other available chicken 8- tubulin genes confirms our earlier suggestion that within an otherwise conserved framework, sequences within two variable region domains serve to define specific &tubulin polypeptide isotypes. The data indi- cate that the c83 gene encodes a unique @-tubulin iso- type which diverges from the dominant neuronal 8- tubulin isotype in 18 of 445 residues (4%). Although the protein coding regions of the ea3 gene are highly homologous to the chicken cS1, c82, c84, and e85 genes previously reported by us, no significant sequence ho- mology with these previously analyzed genes is dis- cernible in the 5’- or 3’-untranslated region sequences, in the intervening sequences, or in the presumptive transcriptional promoter sequences.

The principal structural subunit of eukaryotic microtubules is the protein tubulin, a heterodimer which is comprised of related (but nonidentical) a! and @ polypeptides (see Ref. 1 for review). These cylindrical protein polymers are prominent filamentous structures that provide two essential functions for all dividing cells: 1) construction of mitotic spindles for chromosome separation prior to cell division (2); and 2) in concert with actin filaments and intermediate filaments, es- tablishm~nt of internal cytoarchitecture (3). In addition, mi- crotubules are also utilized for a variety of differentiated cell functions. Not only are they the rigid polymeric elements which establish the highly elongated shape characteristic of neuronal cells, but they also are an integral component of the intracellular transport processes within those asymmetric

* This work was supported by grants from the National Institutes of Health and March of Dimes (to D. W. C. ) . The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ‘ ~ ~ u e r t i s e ~ e n t ~ ’ in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

8 Supported by a National Institutes of Health postdoctoral fellow- ship. Present address: Scripps Clinic and Research Foundation, La Jolla, CA 92037.

11 Recipient of a National Institutes of Health research career development award.

_.

processes (4,5). Microtubules also play a central role in other forms of intracellular transport (6) and in transduction of signals originating at the cell membrane; and, as major struc- tural components of eukaryotic cilia and flagella (3, they represent an important element in some forms of intercellular motility.

Our initial observation that both a- and @-tubulin subunits are encoded in vertebrate organisms by small multigene fam- ilies (Ref. 8; reviewed in Refs. I and 9) rekindled earlier speculations that different microtubule classes might be as- sembled from different tubulin subunits, thereby establishing functionally distinct microtubules (10). In particular, within the chicken genome, we recently documented that at least seven different functional @-tubulin genes are utilized for assembly of microtubules in different cell and tissue types (11,lZ). From these studies, we have concluded that @-tubulin multigene families are utilized in some instances for expres- sion of functionally indistin~ishable polypeptides during al- ternative cell differentiation pathways; whereas in other in- stances, the multiple genes most probably encode polypeptides which confer some unique (although subtle) functional prop- erty to the assembled microtubule (1, 12).

To further investigate the functional properties of the @- tubulin multigene family, we have analyzed @-tubulin synthe- sis in testis. Since during the development and maturation of spermatocytes in testis, microtubules perform most of the specialized cellular ~ n c t i o n s in which microtubules are known to participate, we reasoned that insight into how different @-tubulin genes contribute to differential microtu- bule functions might unfold from such an analysis. To this end, we have determined the complete DNA sequence for c@3, a chicken @-tubulin gene which we show to be the dominant @-tubulin expressed in testis. Like previously characterized vertebrate @-tubulin genes (13-171, this gene is comprised of four exon sequences interrupted by three intervening se- quences at amino acids 19/20,56, and 93. Comparison of the predicted polypeptide sequence for c@3 with four other avail- able chicken @-tubulin sequences confirms our earlier sugges- tion (1,12,16) that within an otherwise conserved framework, sequences within two variable region domains serve to define specific @-tubulin polypeptide isotypes. The data also indicate that the c@3 gene encodes a unique chicken @-tubulin isotype which diverges from the dominant neuronal @-tubulin isotype in 18 of 445 residues (4%). This sequence divergence is con- sistent with the hypothesis that the c@3 polypeptide isotype may be functionally differentiable from other @-tubulin iso- types.

MATERIALS AND METHODS

Isotation of Genomic and cDNA Clones for c&Y--Agenomic segment of chicken DNA containing the c83 gene was isolated as described from an amplified bacteriophage Charon 30 library of chicken DNA

13317

13318 cP3, a ,&Tubulin Gene (18). The tubulin s e q u e n ~ - c o n t ~ n i n g region was previously identi- fied by DNA blot hybridization, and a IO-kilobase pair Hind111 fragment was subcloned into pBR322.

CIones c o ~ e s p o n ~ n g to cDNA copies of the c@3 mRNA were isolated from a X g t l O cDNA library which was constructed from mRNA isolated from immature chicken erythrocytes. Several inde- pendent isolates were identified and plaque-purified, and the cDNA was subcloned into M13 and subjected to DNA sequence analysis. cP3 sequence-containing clones were unambiguously identified by complete nucleotide sequence agreement with the c83 gene, both in the 3'-untranslated region and in the coding sequences.

DNA Sequencing-DNA sequencing was performed using the di- deoxy chain terminator method, as described by Sanger et al. (19). DNA fragments for sequencing were generated by digesting plasmids with a restriction endonuclease followed by Bat31 exonuclease diges- tion (Bethesda Research Laboratories) essentialIy as described by Poncz et al. (20). Following digestion with a second restriction en- zyme, fragments were ligated to the appropriately cut M13 vector (21) and transformed into Escherichia coli JM101. Positive plaques were identified by lifting and probing with an appropriate 3ZP-labeled DNA fragment isolated from the original plasmid. DNA was prepared from individual plaques essentially as described by Sanger et al. (22).

Nucleotide sequence analysis was performed with a VAX/VMS computer using the algorithms developed by Wilbur and Lipman (23) and Goad and Kanehisa (24). Homology within coding regions was calculated using the procedure described by Chan et al. (25) as modified by Perler et ai. (26).

RNA Analysis-RNA was isolated from primary chicken sympa- thetic neuron cultures using the method of Chirgwin et al. (27). Testis RNA was the generous gift of Marc Groudine (Hutchin~on Cancer Center). Poly(A)+ mRNA was prepared by chromatography on oligo(dT)-cellulose. RNA was quantitated by spectrophotometry.

S1 analysis was performed essentially as described previously (17). A 5' probe for determining the transcription initiation site of cp3 was generated by digestion of a purified 691-bp' SphI-PstI fragment of pPG3 with Sau3AI at a site within the first exon. This probe was labeled at 5' ends using polynucleotide kinase (Pharmacia P-L Bio- chemicals) and [ Y - ~ ~ P ] A T P (Amersham Corp.) following treatment with calf intestinal alkaline phosphatase (Boehringer Mannheim). The strand complementary to c@3 mRNA is thus labeled at the site a t position 12 in the cP3 sequence (with the A of the methionine translation initiation codon as nucleotide +I). For analysis, 100 ng of poly(A)+ mRNA was coprecipitated with 3-10 X mol of probe and hybridized for 16 h at 54 "C in the presence of 80% formamide. Following digestion with SI nuclease (Bethesda Research Laborato- ries), protected fragments were detected by electrophoresis and au- tora~ography.

Denaturing gels for RNA blotting were cast from 0.8% agarose containing 2.2 M formaldehyde (28). After electrophoresis, RNA was transferred to nitrocellulose as previously described (11). Radiola- beled DNA probes were prepared using (a-32P]dATP and the random priming method of Shank et al. (29), as previously described (8). All probes were labeled to approximately equal specific activities, with the exception of 82-3' which gave %fold lower activity. Hybridization and washing of RNA blots were performed as described in detail elsewhere (11).

RESULTS

Sequence of the c@3 @-Tubulin Gene-The genomic DNA segment containing the chicken c@3 gene has been isolated by molecular cloning and subcloned into the piasmid pBR322 (18). Although we have previously reported the partial amino acid sequence of the @-tubulin polypeptide which is encoded by this gene (12, 30), we have now utilized the dideoxy chain terminator method (19) to determine the complete DNA sequence of c@3. A schematic diagram of the gene and the sequencing strategy are outlined in Fig. 1. Sequences from all regions were determined on both strands.

The 2697 nucleotides which have been determined are shown in Fig. 2, along with the predicted sequence of the encoded polypeptide. The c@3 gene possesses a structure typ- ical of vertebrate @-tubulin genes, consisting of four protein -___-

I The abbreviation used is: bp, base pair.

Expressed in Testis

- I - - - - - - - - - L - 1

d "- - - A

L - * f - IOObe

FIG. 1. Structure and sequencing strategy of the c@3 tubulin gene. The top part of the figure shows the regions of ppG3 which were subjected to sequence analysis. The intron/exon structure of the genes is indicated (noncoding sequences (thin lines), coding sequences (boM lines)) and corresponds to that previously reported for other vertebrate @-tubulin genes (13-17). The putative transcriptional con- trol element (TATATAA) is marked, as is the polyadenylation signal ATTAAA. Sites for some restriction endonucleases are marked 0, BamHI; ., EcoRI; 13, SacII; V, SphI; 0, Pst; V, BstEII. Schematic representation for two cDNA clones (clones pT3-A and pT3-B) of c@3 is shown beneath the structure of the gene. Actual segments from which DNA sequences were determined are indicated at the bottom of the figure. Sequences determined from cDNA clones are marked with asterisks.

coding exon sequences interrupted by intervening sequences of 90, 76, and 389 bp between codons 19 and 20 and inter- rupting codons 56 and 93, respectively. In addition, the gene is flanked by an appropriate consensus transcriptional control sequence (TATATAA) 76 bases 5' to the translation initiation codon and by an unusual polyadeny~ation signal ( A T T ~ A ) 123 nucleotides 3' to the translation termination codon.

To confirm the major structural features of the 4 3 gene, we have also isolated cDNA clones copied from the mRNA transcribed from c@3. Sequence analysis of two of these (dia- gramed in Fig. l, middle) verifies the intronfexon borders, the actual site of polyadenylation, and the predicted encoded polypeptide sequence within the carboxyl- and amino-termi- nai variable region domains (see below). In all cases, the sequence of the cDNAs corresponds exactly with that deter- mined from the cP3 gene.

To distinguish unambiguously between promoter and 5'- untranslated regions of the gene, the transcription initiation site for 12/33 was determined by using S1 nuclease analysis as described under "Materials and Methods." Specifically, a frag- ment from c83 that begins at nucleotide 12 with respect to the A of the ATG translation initiation codon was labeled at the 5' ends and was used to probe poly(A)+ RNA from chicken testis. After hybridization and digestion with S1 nuclease, the probe fragments protected from digestion as the result of hybridization to c83 RNAs were electrophoresed on a 6% polyacrylamide, 8 M urea gel and detected by autoradiography. The resultant autoradiograms for S1 reactions in the presence (lane 2) or absence (lane I) of added RNA are shown in Fig. 3. Somewhat surprisingly, a heterogeneous cluster of pro- tected fragments was observed, suggesting the presence of multiple transcription initiation sites. Similar heterogeneity was also found using a variety of S1 nuclease conditions employing higher and lower hybridization temperatures and higher levels of S1 nuclease (not shown). The most prominent of these protected fragments was a doublet which mapped 61- 62 base pairs 5' to the ATG initiation codon (Fig. 3, lane 2). In view of these S1 results and the presence of a putative TATA transcriptional control element (TATATAA) 26 nu- cleotides upstream of this apparent major cap site, we con- clude that the dominant transcription initiation site for c@3 lies at this position (marked in Fig. 2 with arrowheads). In addition, these data are most readily consistent with multiple transcription initiation sites for c@3 within -10 bp of the

cP3, a P-Tubulin Gene Expressed in Testis 13319

100

200

300

400

500

600

700

800

900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

2400

2500

2600

2700

13320 ~(33, a &Tubulin Gene Expressed in Testis

M 1 2 11c

9c

76

6i

34

FIG. 3. Identification of the transcription initiation sites of the c83 gene using S1 nuclease mapping. The transcription initiation site(s) of the cs3 gene was determined by an SI nuclease protection experiment. Chicken testis RNA was hybridized to a c83 DNA probe which was 5' end-labeled at position +I2 (with the A of the ATG translation initiation codon taken as +l). Following hybrid- ization, probe fragments resistant to digestion by the single strand- specific nuclease SI were analyzed by electrophoresis and autoradiog- raphy. Lane I , control reaction to which no RNA was added; lane 2, protected fragments following hybridization to 100 ng of testis poly(A)' RNA; lane M, nucleotide markers in base pairs.

major start site (a situation analogous to that previously documented for other eukaryotic promoters such as the SV40 early promoter).

Sequence determined from cDNA clone pT3-B demon- strated that the site of poly(A) addition occurs 14 bp 3' to a ATTAAA polyadenylation signal.

Outside the Protein Coding Region, No Homology Is Re- tained Between c@3 and Other Known @-Tubulin Genes-In light of our previous demonstration of noncoding homology retained between some tubulin genes both within a species (17) and between species (37),* it was of interest to determine whether such noncoding region homologies could be detected for the cj33 gene. Although the coding region sequences are

* K. F. Sullivan and D. W. Cleveland, unpublished data.

highly homologous both at the protein and nucleotide levels (see below), using the homology routines of Wilber and Lip- man (23) and Goad and Kanehisa (24), we have been unable to identify any significant sequence homology,(beyond that found in the TATA transcriptional promoter, consensus RNA splicing, and polyadenylation signals) in noncoding regions between 12/33 and all other vertebrate &tubulin genes whose sequences are known. These nonhomologous regions include the transcriptinal promoter region (5' to the site of transcrip- tion initiation), the 5"untranslated region, the three inter- vening sequences, the 3'-untranslated region, and the 3' gene flanking sequences.

@-Tubulin Encoded by cj33 Represents a Novel @-Tubulin Isotype-In an effort to determine whether the c@3 gene encodes a polypeptide which might define a unique @-tubulin isotype, we have compared the predicted polypeptide sequence with sequences from the four available chicken j3-tubulins (Fig. 4). Inspection of the figure confirms our earlier sugges- tions (1, 12, 16) that within a highly conserved framework, there exist several clusters of amino acid substitutions which constitute the major regions of sequence heterogeneity among the various polypeptides. In particular, the major region of sequence divergence among the known @-tubulin isotypes is found at the extreme carboxy! terminus. Comparison of this isotype-defining domain clearly indicates that the c@3 poly- peptide is significantly divergent from all previously identified chicken @-tubulin polypeptides. For example, whereas most homologous to the dominant chicken neuronal and skeletal muscle isotypes (which are encoded by the sister genes c@l and c@2), the c@3 polypeptide diverges from the neuronal/ skeletal muscle @-tubulin polypeptide in a total of 6 out of 15 residues within the carboxyl-terminal variable region. Overall, c@3 differs from the c@2 isotype a t 4% (18/445) of residue positions.

A similar analysis comparing the cj33 polypeptide with known, complete vertebrate @-tubulin polypeptide sequences (from human (13-15) or porcine (35)) or with all available partial sequences from mouse (12,34), rat (33), or bovine (36) confirms the finding that cj33 represents a novel j3-tubulin isotype.

c@3 Gene Encodes the Dominant Testis &Tubulin Polypep- tide-We have previously determined that although the c@3 gene is expressed at low levels in a variety of cell and tissue types, it is expressed in high levels in chicken testis (11). This finding raised the possibility that cj33 encoded an abundant, perhaps dominant, @-tubulin in testis. We have also identified 3"untranslated region segments specific for the RNA tran- scripts of five of what we believe to be seven total chicken @- tubulin genes (11). Since we can unambiguously distinguish the presence of RNA derived from a sixth gene (c@4') by its characteristic 3500-3700-base size (11) and since efforts from Murphy and Wallis (31,32) have demonstrated that polypep- tides encoded by a seventh @-tubulin gene accumulate exclu- sively in red blood cell precursors, we can now analyze the expression of each of seven @-tubulin genes within the testis.

For this, replicate RNA blots containing aliquots of testis ( T ) or brain (B) RNAs were prepared and hybridized to a @- tubulin coding sequence (labeled @ in Fig. 5) or to gene- specific probes for c@l, c@2, c@3, c@4, and 12/35. After hybridi- zation and washing, the blots were autoradiographed to reveal the patterns which are shown in Fig. 5. Clearly, c@3 RNA is present a t very high levels in testis, whereas it is barely detectable in brain. Conversely, cj32 is the dominant j3-tubulin in brain, but is expressed a t very low levels in testis. cb1,c@4, and c@4' (the cj34' transcript size is marked by the arrow) are undetectable in testis RNA; whereas RNAs from cj35, while

c@3, a @-Tubulin Gene Expressed in Testis 13321

10 I I R E I V B L Q A G Q C G N ~ G A K P I E V I S D E E G ~ ~ G D S D ~ R I N ~ ~ ~ ~ ~ ~ L ~ S ~ S G P F R ~ F ~ ~ ~ ~ S G A ~ ~ G ~ ~ L ~ S ~ W

20 30 40 so 60 70 80 90 100 110 120

I I

I G G

I I T G V sss G L

A GEL I I T

130 140 150 160 170 180 190 200 210 220 230 240 R K E A E S C D C L Q G F ~ T E S U S G M ~ I S K I R E E Y P D R I ~ S V V P S P W S D T V V E P Y N A T L S V B Q L V E N I C F R n g L ~ ~ L N H L V S A T M S O Y I T C L

S S C N C A

M V M

M V I A S

S

250 260 270 280 290 300 310 320 330 340 350 360 R F P G Q L N A D L R K L A V N M V P F P R L B F F M P G P A P L T S R G S ~ ~ ~ D L T ~ ~ C ~ G E G ~ ~ M ~ R G ~ S M ~ ~ E ~ ~ S S ~ ~ I P ~ A V ~ I P P ~

E S D R I E

R E S D R I

D R T AI S A

V E D R T P AI V

370 3 80 390 400 410 420 L l l M S A T F I G N S T A I Q E L F g I S E a F T A M F R R g A F L B W r r G E G ~ ~ ~ S ~ L V S E Y ~ Y ~ A T ~ E E G E - F E E ~ E ~

D Q - G DEAX MY DDE SEW S

AS S F FIG. 4. Comparison of the c@3 polypeptide sequence with all other known chicken &tubulins. The

predicted polypeptide sequence of the c83 gene is shown on the top line. Sequences of other chicken 8-tubulin genes are shown in comparison, explicitly displaying only those residues which differ from the c@3 sequence. Previously identified variable region domains (at the carboxyl terminus beyond residue 430 and near the amino terminus between residues 35 and 57) are boxed. Sequences for the chicken tubulins are taken from Ref. 17 (c81 and c@2), Ref. 16 (~$4) and Ref. 12 (~05) . These comparisons have been previously reviewed by us along with all other known vetebrate @-tubulins in Ref. 12.

detectable, are very minor transcripts with respect to c@3 levels. Correcting for the specific activities and lengths of the various probes employed and assuming that RNA abundance correlates with level of expression, we conclude that the overwhelmingly dominant &tubulin expressed in chicken tes- tis is the c@3 isotype.

DISCUSSION

Together with our previous efforts (12, 16, 17, 49), our present analysis of the c@3 gene has established the sequences of five of what we believe to be seven @-tubulin genes in the chicken genome. One (cpl) of these genes encodes an abun- dant skeletal muscle subunit (11); one (c@2) encodes the dominant neuronal @-tubulin (11); one (c@4) represents a p- tubulin expressed exclusively in neurons (but at levels much lower than c@2 (16, 49)); and one (c@5) is expressed ubiqui- tously except for its absence in neurons, a pattern of expres- sion which is complementary to that of c@4 (49). As we have shown here, 12/33 represents the overwhelmingly dominant 0- tubulin expressed in testis although it is also expressed a t much lower levels in other tissue/cell types (11). The remain- ing two members of the chicken @-tubulin gene family include cp4’ which encodes a @-tubulin which is ubiquitously ex- pressed (11) and c@6 which is expressed dominantly and perhaps exclusively in red blood cell precursors (31, 32).

As we earlier suggested on the basis of partial sequence data (1,16,30), what emerges from inspection of the polypep- tide sequences encoded by each of the five analyzed chicken @-tubulins (Fig. 4) is the realization that within an otherwise highly conserved polypeptide backbone, there exist clusters of heterogeneity which (along with additional single residue substitutions) serve to define four different P-tubulin isotypic sequences encoded by the five isolated chicken genes. The

P P1 P3 P4 P5 T

Y

FIG. 5. @-tubulin polypeptide encoded by 4 3 is the over- whelmingly dominant isotype expressed in testis. Replicate aliquots (1 pg) of poly(A)+ RNA from chicken testis or from brain were electrophoresed on a 0.8% agarose gel containing 2.2 M formal- dehyde and then blotted to nitrocellulose. Blots were then hybridized to 32P-labeled probes prepared from: pT2 (8), a cDNA clone for cj32 which contains the entire coding sequence and which recognizes all known vertebrate &tubulin RNAs (part 8); @1-3’, a 475-bp fragment of the c01 gene which begins 24 bases 3’ to the translation termination codon and extends into the cpl 3’-untranslated region (11) (part 81); j32-3’, a 240-bp fragment containing the 3’-untranslated region of the c82 RNA transcript (11) (part 82); 83-3’, a 2000-bp fragment obtained from the Bal31 sequencing strategy and which begins a t the transla- tion termination codon and contains the entirety of the c83 3’- untranslated region (11) (part 83); 84-3’. a 659-bp fragment of c84 which contains 7 bp of coding and the entire 3”untranslated region of the c04 RNA (11) (part p4); and j35-3’, a 3.3-kilobase pair fragment of the c85 gene which extends from a Sac1 site within the carboxyl- terminal coding region through to a KpnI site in the 3”flanking region (part 85). The transcript derived from c84’ is distinguished by size (arrow). T, testis; B, brain.

most prominent of these variable region domains is located in the extreme carboxyl-terminal sequences beyond residue 430 which display variation a t 76% (13 of 17) of residue positions including insertions, deletions, and terminal addi-

13322 cp3, a p - ~ u b ~ l i n Gene Expressed in Testis

tion of amino acids. A less striking cluster of heterogeneity is localized between residues 33 and 57.

Detailed comparison of the five @-tubulin sequences further reveals that the polypeptides encoded by cpl, cp2, and c@3 are significantly more closely related to each other than they are to either c@4 or c@5 (4% divergence versus >8%). On this basis and in view of the complementary programs of expres- sion of c@4 and c@5, we have suggested that at least two divergent lines of genes (one containing c@4 and c85 and the other containing cp1, cP2, and c@3) encode /3-tubulin within the chicken genome (49).

The significant sequence heterogeneity identified within the chicken @-tubulin gene family and in other vertebrate genomes (12-15, 33-36) rekindles previous suggestions that microtubule differentiation may be achieved (at least in part) by variations in the tubulin subunits themselves (1, 10, 12). Although the failure to detect such properties in in vitro assays (e.g. Ref. 38) or in gene transfection experiments (39) could be interpreted to indicate that no such isotypic distinctions are in fact utilized, direct evidence that @-tubulin isotypes are biochemically distinguished in some instances within cells has been presented by Gard and Kirschner (40) and Edde et dl. (41). Both of these latter groups have documented the differ- entiation-dependent phosphorylation of an electrophoreti- cally distinct isoform of p-tubulin in cultured neuroblastoma cells. Clearly, the presence of cellular mechanisms which distinguish among variant tubulins mandates that such var- iants are in fact utilized for unique roles in cells.

But what are those roles and which of the different @- tubulin polypeptides are utilized for such specialized func- tions? In particular, what can be discerned concerning func- tional aspects of cP3, which we have shown here to encode the overwhelmingly dominant @-tubulin in chicken testis? Testis provides an important model system because during spermatogenesis, microtubules are utilized both for mitotic and meiotic chromosome segregation, for construction of sperm flagella, and for maintenance of internal cytoarchitec- ture. We have found, like the extensiveiy studied invertebrate example in Drosophila (Refs. 42-46; reviewed in Ref. 47), that a single @-tubulin isotype comprises the bulk of the @-tubulin subunits in chicken testis. Since all of the other chicken genes are expressed in testis at only low to undetectable levels (Fig. 5 ) and in light of the elegant work of Raff and co-workers (45) who used a combination of genetic and biochemical methods to document that the single dominant @-tubulin expressed in Drosophila testis must be utilized to construct all classes of testis microtubules, we infer that c@3 must also be m~ltifunctional. We would quickly caution, however, that neither the Drosophila nor the chicken data are sufficient to exclude the possibility in either species that the dominant testis polypeptides (each of which are expressed at low levels in other tissues i l l , 48)) possess some unique (if subtle) functional property.

Acknowledgments-We thank Dr. Marc Groudine for the gift of testis RNA. We also thank Floyd Benko for his tireless technical assistance during the early stages.

REFERENCES 1. Cleveland, D., and Sullivan, K. (1985) Annu. Reu. Biochem. 5 4 , 331-365 2. Inoue, S. (1981) J. Cell Biol. 91,131s-147s 3. Heuser, J. and Kirschner, M. (1980) J. Cell Biol. 86, 212-234 4. Hayden, JI, and Allen, R. (1984) J. Cell Biol. 99, 1785-1793 5. Vale, R, Schnapp, B., Reese, T., and Sheetz, M. (1985) Cell 40,449-454 6. Schliwa, M. (1984) in Cell and Muscle Motility (Shay, J., ed) Vol. 5, pp. 1-

82, Plenum Press, New York 7. Gibbons, I. (19753 in Mokcules andcell Mouement (Inoue, S., and Stephens,

8. Cleveland, D., Lopata, M., MacDonald, R., Cowan, N., Rutter, W., and R. D., eds) pp. 207-232, Raven Press, New York

Kirschner, M. (1980) Cell20,95-105

10. Fulton, C.: and Simpson, P. (1976) in Cell Motili (GoIdman, R., Pollard, 9. Cleveland D. (1983) Cell 34 , 330-332

T., and Rosenbaum, J., eds) pp. 987-1005, CoirSpring Harbor Labora- tory, Cold Spring Harbor, NY

11. Havercroft, J., and Cleveland, D. (1984) J. Cell Bid . 9 9 , 1927-1935 12. Sullivan, K., and Cleveland, D. W. (1986) Proc. Natl. Acad. Sei. U. S. A, ,

83.4327-4381 13. Gwo-Shu~Lee,"., Lewis, S., Wilde, C., and Cowan, N. (1983) Cell 33,477-

14. Lewis, S., Gilmarkin, M., Hall, J., and Cowan, N. (19851 J. Mol. Biol. 182,

15. Gwo-Shu Lee. M.. Loomis. M.. and Cowan. N. (1984) Nucleic Acids Res.

487

11-20 ~I

12,5823-5&36 '

I ~ ~, ~~ ~

16. Sullivan, K., and Cleveland, D. (1984) J. Cell Biol. 9 9 , 1754-1760 17. Sullivan, K., Lau, J., and Cleveland, D. (1985) Mol. CeU. Biol. 5,2454-2465 18. Lopata, M., Havercroft, J., Chow, L., and Cievetand, D. (1983) Cell 32,

19. Sanger, F., Coulson, A., Barrell, B., Smith, J., and Roe, B. (1980) J. Mol. 713-724

20. Poncz, M., Solowiejczyk, D., Ballantine, M., Schwartz, E., and Surrey, S. Biol. 143,161-169

21. Messing, J., and Vieira, J. (1982) Gene (Amst . ) 19,269-276 (1982) Pm. Hat1 Acad. Scz. U. S. A. 79,4298-4302

22. Sanger, F., Nicken, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. s. A. 74,5463-5467 23. Wilbur, W., and Lipman, D. (1983) Proc. Natl. Acad. Sci. U. S. A. 80,726-

730 24. Goad, W., and Kanehisa, M. (1982) Nucleic Acids Res. 10,247-263 25. Chan, S., E iskopu, V, Zeitlin, S., Karathanasis, S., MacKrell, A., Steiner,

D., and &stratiadis, A. (1984) Proc. Natl. Acad. Sei. G'. S. A. 8 1 , 5046- 5050

26. Perler, F., Efstratiadis, A., Lomedico, P., Gilbert, W., Dolodner, R., and Dodgson, J. (19801 Cell 20,555-565

27. Chirgwin, J., Przybyla, A., MacDonald, R., and Rutter, W. (1979) Biochem- istry 18,5294-5299

28. Boedtker, H. (1971) Biochim. Biophys. Acta 240,448-453 29. Shank, P. R., Hughes, S. H., Kung, H.-J., Majors, J. E., Quintrell, N.,

Guntaka, R. V., Bishop, J. M., and Varmus, H. E. (1978) CeU 15,1383- 1.195

30. Suiliian, K. F., Havercroft, J. C., and Cleveland, D. W. (1984) in Molecular &o&y ofthe Cytoskeleton {Borisy, G. G., Cleveland, D. W., and Murphy, D. G., eds) pp. 321-332, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

31. Murphy, D., and Wallis, K. (1983) J. Biol. Chem. 258,7870-7875 32. Murphy, D., and Wallis, K. (1983) J. Biol. Chem. 258,8357-8364 33. Farmer, S. R., Bond, J. F., Robinson, G. S., Mhangkollo, D., Fenton, D.,

and Berkowitz, E. M. (1984) in Molecular Biology of the Cytoskeleton (Borisy, G. G., Cleveland, D. W., and Murphy, D. B., eds) pp. 333-342, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

34. Lewis, S., Lee, G.-S. M., and Cowan, N. (1985) J. Cell Biol. 101,852-861 35. Krauhs, E., Little, M., Kempf, T., Hofer-Warhmek, R., Ade, W., and

36. Little, M., and Luduena, R. (1985) EMBO J. 4,51-56 Ponstingl, H. (1981) Proc. Nut(. Acad. Sci. [I. S. A. 78,4156-4160

37. Cowan. N. J.. Dobner. P. R.. Fuchs. E. V.. and Cleveland. D. W. (1983)

38. 39.

40. 41.

42.

43.

MoL'Celt. BWZ. 3, 1738-1745 . .

Bond, J. F., Fridovic~-Keil, J. I,., Plllus, L., Mulllgan, R C., and Solomon, Swan, J. A,, and Solomon, F. (1984) J, Cell Biol. Q9, 2108-2113

Gard, D. L., and Kirschner, M. W. (1985) J. Cell Biol. 100 , 764-774 Edde, B., Jeantet, C., and Gros, F. (1981) Biochem. Biophys. Res. Commun

Kemphues, K. J.? Raff, R. A., Kaufman, T. C., and Raff, E. C. (1979) Proc.

Kemphues, K. J., Raff, E. C., Raff, R. A,, and Kaufman, T. C. (1980) Cell

F. (1986) Cell 44,461-468

103,1035-1043

Natl. Acad. Set. U. S . A. 76,3991-3995

44. Raff, E. C., Fuller, M. T., Kaufman, T. C., Kemphues, K. J., Rudolph, J.

45. Kemphues, K. J., Kaufman, T. C., Rsff, R. A., and Raff, E. C. (1982) Cell

46. Kemphues, K. J., Raff, E. C., and Kaufman, T. C. (1983) Genetics 106 ,

21,445-451

E., and Raff, R. A. (1982) Cell 28,33-40

3 1,655-670 IAiaKC:

47. Raff, E. C. (1984) J. Celt Biol. 9 9 , 1-10 48, Natzle, J. E., and McCarthy, B. J. (1984) Deu. Biol. 104,187-198 49. Sullivan, K. F., Havercroft. J. C., Machlin, P. S., and Cleveland, D. W.

I,_" """

(1986) Mol. Cell. Blol., in press