Microtubule Architecture Specified by a β-Tubulin Isoform

See allHide authors and affiliations

Science  03 Jan 1997:
Vol. 275, Issue 5296, pp. 70-73
DOI: 10.1126/science.275.5296.70


In Drosophila melanogaster, a testis-specific β-tubulin (β2) is required for spermatogenesis. A sequence motif was identified in carboxyl termini of axonemal β-tubulins in diverse taxa. As a test of whether orthologous β-tubulins from different species are functionally equivalent, the moth Heliothis virescens β2 homolog was expressed in Drosophila testes. When coexpressed with β2, the moth isoform imposed the 16-protofilament structure characteristic of that found in the moth on the corresponding subset of Drosophila microtubules, which normally contain only 13-protofilament microtubules. Thus, the architecture of the microtubule cytoskeleton can be directed by a component β-tubulin.

In eukaryotic cells, microtubules form diverse structures that are used for many different functions. Within each microtubule array, there are two levels of supramolecular organization: the architecture of each individual microtubule, determined by the number and arrangement of protofilaments, and the overall morphology of the microtubule array, for example, an axoneme or a spindle. Morphogenesis of each structure depends both on interactions between α- and β-tubulin heterodimers and on interactions between tubulins and other proteins. In vertebrate β-tubulins, isotype-defining variable regions (in particular the COOH-terminus), which have diverged among different isoforms in a gene family but are conserved in orthologs from different species, have been postulated to have an important role in conferring the functional specificity of each class of isoform (1). We previously demonstrated the validity of this hypothesis by showing that the unique COOH-terminus of the Drosophila melanogaster testis-specific β2-tubulin isoform is required for tissue-specific functions, including morphogenesis of the motile axoneme (2, 3). Spermatogenic-specific microtubule functions cannot be provided by β3, another Drosophila β-tubulin isoform normally used during differentiation of a variety of somatic cells (3, 4).

Vertebrate β-tubulin orthologs in different species are conserved in structure and have similar expression patterns, suggesting that they perform similar functions (1). If this model is true for other groups of organisms of similar evolutionary relationship, then the β2 ortholog from another insect should be better able to function in the Drosophila male germ cells than the paralogous β3 isoform. A cDNA from the moth Heliothis virescens was reported that represents the gene for a testis-specific β-tubulin (Hvβt) whose expression pattern suggested it to be the moth β2 ortholog (5). We sequenced the Hvβt clone and compared the predicted amino acid sequence with Drosophila β-tubulins. The COOH-termini of Hvβt and β2 were more similar to each other than to the other Drosophila β-tubulins and exhibited the same relative similarities to the other Drosophila isoforms (Fig. 1A), substantiating that Hvβt is orthologous to β2. However, unlike the highly conserved orthologs in vertebrate β-tubulin families [96 to 99% identity (1)], Hvβt in its entirety was equally distant from all of the Drosophila β-tubulins (approximately 80% identity, 90% similarity). To confirm the relation between β2 and Hvβt, we therefore compared other axonemal β-tubulins with isoforms not used in motile axonemes. Axonemal β-tubulins from diverse taxa have a motif consisting of the consensus sequence EGEF followed by three acidic residues present at the same position in the COOH-terminus (Fig. 1B). Presence of the sequence motif does not preclude function in other kinds of microtubules, but the motif is absent in many β-tubulins that are not used for motile axonemes. Conservation of the axoneme motif throughout eukaryotic phyla is consistent with the hypothesis that assembly of a motile axoneme imposes structural constraints that limit evolutionary divergence of β-tubulins (6, 7). Presence of the motif supports the conclusion that Hvβt is the ortholog for β2. However, the sequence divergence and the functional differences we report below between moth and fly testis β-tubulins demonstrate that the conservation of the β-tubulin gene families in vertebrates is not the general rule for β-tubulin families in other groups with similar times of evolutionary separation. Spermatogenesis in the moth is similar to the process in Drosophila and uses similar sets of microtubules (8). To test whether the testis β-tubulin orthologs have equivalent functional properties, we generated transgenic stocks that express Hvβt in the Drosophila male germ cells (Fig. 2). We examined Hvβt function when it was the sole β-tubulin in Drosophila spermatids and when it was coexpressed with β2. Contrary to our model, Hvβt failed to support any microtubule assembly at all in homozygous β2null male flies (9). Furthermore, when it was coexpressed with β2, Hvβt poisoned all microtubule-mediated processes (including meiosis, spermatid alignment, nuclear shaping, mitochondrial derivative elongation, and axoneme assembly), such that transgenic males in which Hvβt composed more than about 6% of the total β-tubulin pool were sterile (10). The severity of the dominant phenotype reflected the contribution of the moth protein to the germline tubulin pool; additional copies of the β2 gene ameliorated defects in all classes of microtubules.

Fig. 1.

(A) Percent identity of the COOH-terminal sequences of Hvβt and Drosophila (Dm) β-tubulins. (B) An axoneme motif (underlined) in the COOH-termini of β-tubulins used in motile axonemes: plus sign, used in motile axonemes; minus sign, not used in axonemes. Sequences shown begin at residue 431 in most β-tubulins, following a highly conserved region penultimate to the COOH-terminus. Top group, metazoan β-tubulins with well-defined expression patterns (14). Hum, human; Chi, chicken; C.e. mec7, C. elegans touch neuron-specific isoform (11). Middle group, β-tubulins from fungal species in which motile axonemes are not made. S. cer., Saccharomyces cerevisiae; S. pombe, Schizosaccharomyces pombe; Asp, Aspergillus nidulans (15). Bottom group, sequences from ciliated or flagellated protists with only a single β-tubulin (16). Para, Paramecium tetraurelia; Chlamy, Chlamydomonas reinhardtii; Euplo, Euplotes octocarinatus; Giardia, Giardia lamblia. Complete sequences are published and available in databases, except for Drosophila β4 (17). The Hvβt sequence is available under GenBank accession number U75868. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

Fig. 2.

Expression of the Heliothis testis-specific β-tubulin in Drosophila males. The Hvβt coding sequence was expressed in the postmitotic male germ cells under the control of regulatory sequences from the Drosophila β2 gene (18). Hvβt was identified as a novel testis β-tubulin that migrated in two-dimensional gels at the electrophoretic position predicted from the protein sequence. (A) Testis tubulins in a sterile male with one copy of the Hvβt transgene and two copies of the wild-type β2 gene. Left, autoradiogram showing incorporation of 35S-labeled methionine into newly synthesized testis proteins. Right, protein immunoblot of the same gel showing the relative contribution of Hvβt and β2 to the testis tubulin pool. (B) Protein immunoblot showing testis tubulins in a sterile male with one copy of the Hvβt transgene in a β2null background. Immunoblots were probed with antisera to β-tubulin, α-tubulin, and actin. In addition to Hvβt (Hv) and endogenous β2 expressed in the postmitotic male germ cells, positions of β1-tubulin, expressed in earlier spermatogenic stages, α-tubulin (α), and actin (A) are also indicated. β3-Tubulin, expressed in somatic testis cells (3, 4), is visible as a small spot above β1. Quantitation of [35S]methionine incorporation showed that Hvβt and β2 synthesis is proportional to the gene copy number. Stability of Hvβt is the same in the absence or presence of β2.

Figure 3 illustrates the profound failure of cytoplasmic microtubule function (A and B) and axoneme assembly (C to L) that resulted from coexpression of Hvβt with β2. Wild-type axonemes in sperm of moths and flies have the morphology typical for insects, consisting of the highly conserved pattern of nine doublet microtubules surrounding a central pair of two singlet microtubules, plus an additional outer circle of nine singlet accessory microtubules (Fig. 3, C to E). Another unique feature of the insect axoneme is that in the mature sperm, the lumen of each central pair and accessory microtubule contains a filament (2, 3) that in cross section appears as an electron-dense structure in the center of each microtubule (Fig. 3, C and E). The major morphological difference between fly and moth axonemes is that the accessory microtubules in the fly are the same diameter as the central pair, whereas the accessory microtubules in the moth are larger than the central pair (most clearly seen in Fig. 3D in the immature moth axoneme before the luminal structure is present). Tannic acid staining revealed that the difference in diameter reflects a difference in protofilament (pf) number: Moth accessory microtubules are 16-pf compared with 13-pf in the fly (Fig. 3, H to J and L).

Fig. 3.

Spermatogenesis in wild-type D. melanogaster and H. virescens males and in transgenic Drosophila males in which Hvβt is coexpressed with endogenous β2. In (A) and (B) light micrographs of orcein-stained testes show spermatid alignment and nuclear shaping (microtubule-mediated processes). Bar in (A) = 20 μm. (A) Spermatids in a wild-type Drosophila male. Nuclei are shaped and aligned at the tip of the developing bundle (arrow). (B) Spermatids in a sterile transgenic male in which Hvβt constituted 10 to 15% of the β-tubulin pool. Nuclei (arrows) are not shaped and spermatids are not aligned. Spermatid elongation is defective, reflecting failure of axoneme assembly and mitochondrial derivative elongation. Axoneme ultrastructure is shown in (C) to (L) (19). Bars = 50 nm [in (E) for panels (C) to (E); in (G) for panels (F) and (G); and in (K) for panels (H) to (K)]. (C) Mature axoneme with wild-type morphology from a fertile transgenic male in which low amounts of Hvβt were present. (D) Immature Heliothis axoneme; accessory and central pair microtubules do not yet contain luminal structures. (E) Mature Heliothis axoneme. (F to G) Aberrant axonemes in sterile transgenic males in which Hvβt constituted 8 to 10% of the β-tubulin pool. Partial axoneme with accessory microtubules in the process of assembly is shown in (F); formation of one is abnormal (arrow). Fragmented axoneme in a mature spermatid is shown in (G). One accessory microtubule is of large diameter (arrow). In (H) to (L) testes are stained with tannic acid to display the microtubule pf number (20). (H) Immature Drosophila axoneme. Accessory microtubules are completed on the left side of the axoneme but are in various stages of assembly (arrows) on the right side (21). (I) Mature Drosophila axoneme. (J) Immature Heliothis axoneme. (K) Partial axoneme in a mature spermatid from a sterile transgenic male in which Hvβt constituted 10 to 15% of the β-tubulin pool. The central pair and three of the accessory microtubules are 13-pf, but one of the accessory microtubules is 16-pf (arrow). (L) Enlarged views showing pf architecture. Hv-C, 13-pf, from (J); Hv-A, 16-pf, from (J); Hz-A, 16-pf, from a mature axoneme; Dm-A (left), 13-pf, from (H); Dm-A (middle), 13-pf, from (I); Hvβt, 16-pf accessory microtubule in transgenic male, from (K). Abbreviations: A, accessory microtubule; C, central pair microtubule; Hv, H. virescens; Dm, D. melanogaster; Hz, Heliothis zea (spermatogenesis is identical in H. virescens and H. zea). All panels are at the same magnification (21).

In sterile transgenic males, axonemal microtubules were assembled, but axonemes in most spermatids were fragmentary (Fig. 3, F, G, and K). In many of the abortive axonemes, one or more of the accessory microtubules were larger than normal (Fig. 3, G and K), similar to moth accessory microtubules. The larger accessory microtubules in transgenic males were 16-pf instead of 13-pf (Fig. 3, K and L), just as in normal moth axonemes. The unusual protofilament architecture was specific to the axoneme accessory microtubules; no other microtubules in transgenic males had this architecture. The moth protein thus imposed the moth-specific accessory microtubule architecture on the equivalent structures in fly cells, even though it composed only a small percent of the total β-tubulin pool.

How did the moth protein specify protofilament architecture of the accessory microtubules? One possibility would be that in moth spermatids, Hvβt is dedicated to assembling only the 16-pf accessory microtubules. However, this is unlikely, given that, like β2 in the fly, Hvβt is the predominant isoform in the moth testis (5). The templating process for the accessory microtubules in the insect sperm axoneme is unique: Each accessory microtubule is initiated as an “outgrowth” from the B-tubule of the associated doublet (8) (Fig. 3H). Because Hvβt forced assembly of 16-pf accessory microtubules in fly cells (but not of other microtubules in either fly or moth), the moth protein must form a unique interaction with other components specific to this assembly mechanism, even in Drosophila cells. We postulate that although conserved COOH-terminal sequences are required for fundamental axoneme-forming functions, sequences elsewhere in the Hvβt protein are responsible for specification of accessory microtubule architecture, as well as for the dominant phenotype in Drosophila spermatids. Thus, sequences other than the COOH-terminus are also likely to play species-specific roles in spermatogenesis. There has been no previous demonstration of microtubule architecture being specific to a particular tubulin. Expression of the Caenorhabditis elegans mec7 β-tubulin is associated with 15-pf touch neuron microtubules (11), whereas other microtubules in the nematode are 11-pf, but the initiation mechanism is not known. In Drosophila, incorporation of β-tubulins into microtubules of differing pf number depends on the cellular context: Specialized 15-pf microtubules that function in wing maturation (12) contain β1 and β3, isoforms that in other cells give rise to typical 13-pf microtubules (4). The ability of heterologous β-tubulins to assemble into multiple pf arrangements in vitro and into many different structures in vivo provides additional evidence that microtubule architecture can be determined by factors extrinsic to the component tubulins (7). For example, the pf number of microtubules assembled in vitro can be controlled by nucleating conditions (13). Our data demonstrate that the templating machinery can be directed by a specific tubulin subunit.

Identification of a common structural feature shared by many axonemal β-tubulins provides evidence for strong conservation of features of tubulin structure required for assembly of the motile axoneme. Nonetheless, the finding that microtubule architecture can be intrinsic to the β-tubulin primary sequence demonstrates selection for a previously unknown control over cellular microtubule assembly. Many studies have shown that transcription factors and other regulatory proteins retain the ability to function in heterologous systems across long phylogenetic distances. Our functional test of Hvβt in the Drosophila testis reveals that the cytoskeletal proteins that carry out the instructions of the regulatory genes may have acquired much more stringent species-specific restrictions on function.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
View Abstract

Navigate This Article