abnormal spindle


REGULATION

Protein Interactions

Abnormal spindle protein, Asp, and the integrity of mitotic centrosomal microtubule organizing centers

The product of the abnormal spindle (asp) gene is an asymmetrically localized component of the centrosome during mitosis, required to focus the poles of the mitotic spindle in vivo. Removing Asp protein function from Drosophila embryo extracts, either by mutation or immunodepletion, results in loss of their ability to restore microtubule-organizing center activity to salt-stripped centrosome preparations. This is corrected by addition of purified Asp protein. Thus, Asp appears to hold together the microtubule-nucleating gamma-tubulin ring complexes that organize the mitotic centrosome (do Carmo Avides, 1999).

The microtubule-nucleating capacity of the animal cell centrosome requires a ring-shaped complex of proteins associated with gamma-tubulin present within the pericentriolar material (PCM). The PCM contains lattice-like structures in which gamma-tubulin has been found associated with pericentrin. It is not known how the properties of the PCM might change during mitosis specifically to nucleate spindle microtubules (do Carmo Avides, 1999).

Asp is a 220-kD microtubule-associated protein (MAP) found at the poles of mitotic spindles in the syncytial embryos of Drosophila melanogaster. It has consensus phosphorylation sites for p34cdc2 and mitogen-activated protein kinases and putative binding domains for actin and calmodulin. Mutations in asp result in abnormal spindle morphology leading to mitotic arrest or to a high frequency of meiotic nondisjunction. Because the mitotic defects of asp mutants are best studied in the larval central nervous system, its distribution in cells of whole-mount preparations of this tissue was examined more carefully. Asp becomes and remains associated with the centrosome throughout mitosis. At telophase, it migrates to microtubules on the spindle side of both daughter nuclei and is not associated with the centrosome in interphase cells. At all mitotic stages from prophase to anaphase, the Asp protein is asymmetrically localized around the gamma-tubulin in the pericentriolar material, where it appears to form a hemispherical cup contacting the spindle microtubules (do Carmo Avides, 1999).

In several asp mutations, Asp does not immunostain at the spindle poles. The majority of mitotically arrested asp cells have bipolar spindles with broad unfocused poles of microtubules. In a small proportion of cells, one pole is sufficiently disorganized so that the spindle appears monopolar. The gamma-tubulin is not present within a well-organized centrosome in these cells, but is found in dispersed clumps at the spindle poles. This suggests that Asp might be required to maintain the structure of the centrosomal microtubule-organizing center (MTOC) during mitosis (do Carmo Avides, 1999).

To confirm that Asp is a centrosomal protein, centrosomes were partially purified from syncytial Drosophila embryos undergoing their rapid nuclear division cycles. Immunoblotting experiments indicate that this preparation is enriched in both Asp and gamma-tubulin. Moreover, these two proteins colocalize by immunofluorescence at in vitro organizing centers for rhodamine-labeled microtubules. In contrast to observations in vivo, Asp is found in all of the in vitro MTOCs and is distributed symmetrically. This suggests that the extract is in a mitotic-like state, probably due to the dominant effect of the active mitotic protein kinase p34cdc2. It also implies that the asymmetric localization seen in intact cells requires that microtubules make contact with chromosomes to form a spindle. If Asp localizes to the 'outside' of the centrosome, would antibodies to Asp interfere sterically with microtubule nucleation in an in vitro assay? The centrosome preparation was therefore incubated with either affinity-purified anti-Asp or control rabbit immunoglobulins before addition of rhodamine-labeled tubulin. The number of asters of microtubules formed decreases in proportion to the concentration of antibody. However, the number of gamma-tubulin-positive bodies remains constant, suggesting that the antibody is blocking microtubule nucleation rather than disrupting the centrosomes (do Carmo Avides, 1999).

Extraction of centrosomes with KI effectively destroys their microtubule-nucleating activity, leaving a centrosomal scaffold. Ability to nucleate microtubules can be restored to such KI-extracted centrosomes by incubation with the soluble fraction of a Drosophila embryonic extract. Although gamma-tubulin and its ring complex (gammaTuRC) are required to rescue the aster-forming ability of KI-extracted centrosomes, they are not sufficient and have to be supplemented by a high molecular weight microtubule-associated factor speculated to be pericentrin. To test whether Asp protein might be required to reconstitute MTOCs from KI-extracted centrosomes, Asp was immunodepleted from a soluble embryonic extract under conditions where several other centrosomally associated proteins, including gamma-tubulin, CP190, KLP61F, and Polo, were not removed. This immunodepleted extract was unable to restore the ability of KI-extracted centrosomes to nucleate microtubules into asters. However, many linear arrays of microtubules were seen, suggesting that the Asp-depleted extract provided microtubule nucleation ability, but that it was not organized into discrete centers (do Carmo Avides, 1999).

In contrast to the soluble extract of wild-type embryos, the equivalent soluble fraction prepared from asp-derived mutant embryos is unable to restore the ability of KI-extracted centrosomes to nucleate asters of microtubules. Whether addition of purified Asp protein could restore this ability to the mutant embryo extract was investigated. Asp copurifies with microtubules from Drosophila embryos but is not released by concentrations of NaCl known to remove most MAPs. Thus, it seemed that if KI extracts Asp from the centrosomes, it probably would do so from such purified microtubules. NaCl-washed microtubule preparations were extracted with 2 M KI, and a 220-kD protein recognized by antibodies to Asp was found to be the main component of the resulting supernatant fraction. This purified Asp fraction was added to the soluble extract prepared from asp-derived embryos and found to restore the ability of KI-extracted centrosomes to nucleate asters of microtubules (do Carmo Avides, 1999).

Thus, it appears that both Asp and gammaTuRC are required to restore microtubule nucleating activity to centrosome scaffolds. These data do not rule out the possibility that in addition to Asp, other high molecular weight proteins are required for this function. Because Asp neither copurifies with the gammaTuRC nor coimmunoprecipitates with gamma-tubulin, it is unlikely to have a direct role in the nucleation process. Rather, the consequences of loss of Asp function upon spindle poles in vivo and upon MTOCs in vitro suggest that it is required to organize the gammaTuRC within the PCM to form a nucleating center for microtubules at mitosis. In asp mutants, a spindle can still form, most likely reflecting the known ability of mitotic chromatin and motor proteins to organize a bipolar spindle in the absence of centrosomes. However, one consequence of the disorganized centrosomes and spindle poles is that the cells arrest in a metaphase-like state, suggesting that the spindle integrity checkpoint has been activated. Asp protein function is likely to be modified later in the mitotic cycle, since it was observed to associate with the microtubules of the telophase spindle, a property consistent with its purification as a MAP. However, the lack of any obvious association with microtubules during interphase suggests that its properties have to be modulated, possibly by phosphorylation, during entry into mitosis in order to activate its essential role in maintaining the coherence of the centrosome at the spindle poles (do Carmo Avides, 1999).

Mutation of a Drosophila gamma tubulin ring complex subunit encoded by discs degenerate-4 differentially disrupts centrosomal protein localization

The Drosophila gene discs degenerate-4 (dd4) has been cloned and found to encode a component of the gamma-tubulin ring complex (gammaTuRC) homologous to Spc98 of budding yeast. This provides the first opportunity to study decreased function of a member of the gamma-tubulin ring complex, other than gamma-tubulin itself, in a metazoan cell. gamma-tubulin is no longer found at the centrosomes but is dispersed throughout dd4 cells and yet bipolar metaphase spindles do form, although these have a dramatically decreased density of microtubules. Centrosomin (CNN) remains in broad discrete bodies but only at the focused poles of such spindles, whereas Asp (abnormal spindle protein) is always present at the presumptive minus ends of microtubules, whether or not they are focused. This is consistent with the proposed role of Asp in coordinating the nucleation of mitotic microtubule organizing centers. The centrosome associated protein CP190 is partially lost from the spindle poles in dd4 cells supporting a weak interaction with gamma-tubulin, and the displaced protein accumulates in the vicinity of chromosomes. Electron microscopy indicates not only that the poles of dd4 cells have irregular amounts of pericentriolar material, but also that they can have abnormal centrioles. In six dd4 cells subjected to serial sectioning, centrioles were missing from one of the two poles. This suggests that in addition to its role in nucleating cytoplasmic and spindle microtubules, the gammaTuRC is also essential to the structure of centrioles and the separation of centrosomes (Barbosa, 2000).

The major microtubule organizing center (MTOC) in animal cells is the centrosome, which nucleates the slowly growing minus ends of microtubules allowing the plus ends to extend into the cytoplasm. In most animal cells, centrosomes are essential for definition of the interphase MT arrays, for determination of cell polarity, and for the formation and function of the spindle in mitosis. There are two main components of the centrosome: a pair of centrioles comprising cylinders of nine triplet microtubules and the pericentriolar material (PCM) that appears to provide nucleation centers for cytoplasmic and spindle microtubules. Little is known about the organization of the PCM, although both pericentrin and gamma-tubulin have been described as forming a protein complex organized into a lattice like structure. gamma-tubulin is a conserved member of the tubulin family found at MTOCs, including animal cell centrosomes, and the equivalent organelles of yeasts, the spindle pole bodies (SPBs). The gamma-tubulin of S. cerevisiae forms a 6S complex with Spc98p and Spc97p, associated with both the inner and outer plaques of the SPB. Temperature sensitive mutants of its structural gene, tub4, show defects in microtubule nucleation at the newly formed SPB as well as in the assembly of a mitotic spindle. The SPC98 gene was identified as a dosage-dependent suppressor of the tub4-1(ts) allele. Its gene product appears essential for mitotic spindle formation because cells harboring the temperature sensitive allele spc98-1 or over expressing wild-type protein, duplicate and separate their SPBs but form a defective mitotic spindle. The gene encoding the other main component of this complex, SPC97, was isolated as a suppressor of the spc9-2(ts) mutant. Its temperature sensitive alleles show phenotypes similar to tub4 and spc98 mutants as well as defects in SPB duplication. Spc98p docks the Tub4p complex to the inner plaque of the SPB through the N terminus of Spc110p, a protein that forms a bridge between the inner and central plaques. The Tub4 complex is formed in the cytoplasm. It is transported to the nucleus via the nuclear localization signal (NLS) present in Spc98p. Spc98p at the inner plaque of the SPB is phosphorylated in a cell cycle-dependent manner, whereas Spc98p in the outer plaque does not appear to undergo such modification (Barbosa, 2000).

In higher eukaryotes gamma-tubulin occurs in a 25S-32S complex that has been shown by electron microscopy to have a ring shape leading to the name gamma-tubulin ring complex (gammaTuRC). The Xenopus complex comprises seven proteins: alpha-, beta-, and gamma-tubulin, and additional proteins of 75, 109, 133, and 195 kD. The 109 kD Xenopus protein (Xgrip109) is a homolog of yeast Spc98p. It interacts directly with gamma-tubulin and is essential for microtubule nucleation. The major human and Drosophila gammaTuRCs have very similar protein profiles. The 100 and 101 kD human proteins hGCP2 and hGCP3 correspond to Spc97p and Spc98p, respectively. In Drosophila, a second smaller 240 kD gamma-tubulin complex has been described comprising only gamma-tubulin and the Spc97/98 homolog Dgrip84 and Dgrip91. It is proposed that this is assembled into the complete 3 MDa gamma-TuRC, which contains multiple copies of the heterotrimer plus ancillary proteins (Barbosa, 2000 and references therein).

A model of the gammaTuRC in Drosophila suggests that the complex is assembled in the cytoplasm from the heterotrimer subunits and recruited onto the centrosome, where it nucleates microtubules. In mitosis it functions in concert with Asp, or a protein with equivalent function in other organisms, to organize the spindle microtubules (Avides, 1999). This implies that the centrosome might be essential for MT nucleation and therefore for the formation of spindles. However, in the female meiotic divisions of Drosophila the spindles form in the absence of centrioles and without detectable concentration of gamma-tubulin at the poles. Moreover, loss of centrosomes has been observed from the spindle poles in the syncytial embryos of several mutants, and although these generally lead to the accumulation of mitotic defects, several rounds of mitosis can take place on such spindles. These observations, together with the ability to build spindles without centrosomes in vitro that are able to undertake metaphase and anaphase, has been taken to mean that centrosomes might be dispensable in the formation of a functional spindle in some systems. Recent observations that a functional spindle can still form in mammalian cells after laser ablation of the centrosomes now reinforce this idea (Barbosa, 2000 and references therein).

Because dd4 encodes a component of the gamma-tubulin ring complex, the organization of the mitotic spindle and its poles were examined in cells from the central nervous system of dd4 mutants. The localization of gamma-tubulin was examined in relation to centrosomin (CNN), another component of the Drosophila centrosome. In mitotic cells from wild-type brains, these two proteins colocalize to the two centrosomes. In cells of all three dd4 mutant alleles, gamma-tubulin staining can still be detected; however, it is no longer found in a well-defined body but rather is distributed throughout the cell. In contrast, distinct CNN-containing bodies can be seen in every mitotic cell. However, whereas wild-type cells invariably contained two such bodies (the functional centrosomes), some mutant cells contained only one body stained with CNN, while others contained two (Barbosa, 2000).

In cells stained to reveal the spindle microtubules, the CNN-containing bodies always appear to be associated with a microtubule organizing center. In wild-type cells, these are spherical centrosomes at the spindle poles and astral microtubules. In the dd4 mutant cells, the CNN staining bodies are often less tightly defined structures and astral microtubules are not seen. In some cells the CNN containing body appears to have fragmented, and a satellite body can be seen near the main pole. In those cells having only one CNN-staining body, prominent arrays of microtubules extend between this pole and the chromosomes. Conversely, microtubules extending from the chromosomes to the pole lacking the CNN body exhibit reduced staining, and in some cells this pole appears not to have organized microtubules (Barbosa, 2000).

In addition to the gamma-tubulin ring complex, the Asp protein is also known to be required to nucleate asters of microtubules (Avides, 1999). In wild-type mitotic cells, the Asp protein is found on the face of the centrosome that makes contact with spindle microtubules. This close juxtaposition of gamma-tubulin and Asp is no longer seen in dd4cells in which the gamma-tubulin is dispersed, but Asp maintains a punctate distribution. This punctate staining can be clustered around the spindle poles or clustered in one area and scattered throughout the remaining part of the cell, but is never diffuse as is gamma-tubulin. Immunostaining to reveal microtubules shows that Asp protein is always found at the poles of bipolar spindles, either as a well organized body, but more usually in clustered aggregates. In spindles that had only one focused pole, individual bundles of the microtubules could be seen to extend both from this focus and from small punctate caps of Asp at the unfocused pole towards the chromosomes (Barbosa, 2000).

CP190 is an abundant protein which, together with its partner CP60, associates with mitotic centrosomes in Drosophila cells. Although frequently used as a marker to follow centrosome behavior, its function remains unknown. In wild-type cells, CP190 is found associated with the centrosomes at the spindle pole. In the weakest allele, dd4S, some of the antigen is lost from the poles and appears in the central part of the spindle in the region occupied by the mitotic chromosomes. In the stronger alleles, only a small proportion of CP190 remains at the poles; the remainder is clustered around the chromosomes. Thus, intact gamma-tubulin ring complex appears to be required at the poles in order to maintain the polar association of the CP190 protein (Barbosa, 2000).

In addition to being present in the PCM, gamma-tubulin is also found within the centrioles themselves and appears to be required for centriolar function. To determine whether dd4 mutants showed any irregularity of centriolar structure, the ultrastructure of centrosomes was examined by electron microscopy of serial sections of cells from the larval central nervous system. These data consist of complete sets of serial sections through some four wild-type prometaphase cells with chromosomes undergoing congression, one wild-type cell at metaphase, and six dd41mutant cells that are in a metaphase-like state. The centrosomes have well defined centrioles showing typical arrays of triplet microtubules surrounded by the electron-dense PCM. Arrays of spindle microtubules extend from these centrosomes towards condensed chromosomes. These microtubules occur in bundles that make contact with well defined kinetochore plates on the chromosomes. All of the dd4 cells examined had two opposing poles, generally broad, but secondary poles were often present. Centrioles were present only at one of the two main poles in all six of the cells, but there were up to four of them; they varied in size and were not consistently arranged as perpendicular pairs. The amount of PCM varied from one pole to another showing no correlation with the pattern of centrioles. The density of spindle microtubules in dd4 cells is strikingly reduced compared to the wild-type spindle, consistent with the impression gained from immunostaining. The spindle microtubules make poor contact with the mitotic chromosomes and, in contrast to wild-type cells, it was difficult to see organized plate-like kinetochores on dd4 chromosomes (Barbosa, 2000).

It is of interest to compare phenotypes of mutations in dd4 with mutations in the gamma-tubulin genes. Drosophila has two genes for gamma-tubulin; the one at 23C is expressed in a variety of tissues including brains, imaginal discs and testes, whereas expression of the second at 37C is restricted to ovaries and embryos. Like the dd4 mutants, cells from gamma-tub23C brains display abnormally high levels of chromosome condensation, spindles with defective or absent poles, and polyploidy. However, whereas the mitotic index of dd4 cells is dramatically elevated, the mitotic index of gamma-tub23C cells is reduced relative to wild-type and anaphase figures are very rare. The reasons for these differences are not clear. It is as though reduction in levels of the 23C gamma-tubulin lead to a limited numbers of chromosome duplication cycles in the absence of mitosis, but this is followed by interphase cell cycle arrest. In contrast, the dramatic increase in metaphase figures in dd4 mutants resembles a more typical spindle checkpoint arrest. The frequency of anaphases, not dissimilar in total number to wild type, suggests that cells can evade this checkpoint at some frequency, as has been described in several organisms. The phenotypes of various allelic combinations suggest that the strongest allele, dd41, is not completely amorphic by genetic criteria. Western blotting indicates that there is 80%-90% reduction of the Dgrip91 protein in this allele. Although it is uncertain whether this residual protein has any function, this and the genetic observations suggest there may be some residual function of the gamma-TuRC even in the dd41 mutant (Barbosa, 2000).

It is generally thought that the late survival of larvae with extreme mitotic defects reflects perdurance of maternal contribution to the oocyte from the heterozygous mother, as has been shown for other cell-cycle genes. In the case of dd4, this assumption has been challenged by the report of normal oocytes from homozygous dd41 mitotic ovarian clones arising in heterozygous females. However, preliminary observations of embryonic development in eggs produced from allelic combinations weak enough to give viable escaper females (dd4S/dd4S and dd4S/dd43) do indicate that there is a vital maternal contribution to the oocyte: such eggs appear to have parental DNA but they fail to undergo any development. These observations suggest that if mothers carrying weak enough allelic combinations to be compatible with survival to adulthood cannot build a viable egg, then either the observations of no maternal effect are in error, or the observed clones have sufficient perdurance of the wild-type product to build eggs indistinguishable from normal heterozygotes (Barbosa, 2000).

Together, Dgrip84, Dgrip91, and gamma-tubulin form the three major components of the gamma-TuRC and are homologous to the budding yeast proteins Spc97, Spc98, and Tub4. Genetic and molecular studies show interactions between these genes in budding yeast, and their requirement for SPB structure, duplication, and separation. Interactions between members of this complex and other components of the SPB and spindle are only beginning to be understood. Spc98, for example, binds to the N-terminal region of Spc110p, a coiled-coil protein that spans the inner and central plaques of the SPB. The calmodulin binding C terminus of this protein contacts the central plaque and the N-terminal region, the inner plaque. Thus Spc98 might form an essential link between Spc110 and the spindle microtubules that emanate from the inner plaque and the defective spindle structures seen in spc98 mutants may be a direct consequence of defects in this interaction. The phenotypes of spc98 mutants thus have some parallels with dd4 mutants in abnormal spindle microtubule density and organization, and it will be of interest to determine whether Dgrip91 has similar interactions with specific components of the centrosome (Barbosa, 2000).

The more drastic disruption of purified preparations of centrosomes with the salt KI in vitro removes a set of proteins, including the gamma-TuRC, CP60, CP190, CNN, and Asp, thus destroying their ability to organize microtubules. The salt treatment appears to leave behind unidentified core centrosomal components, since the structure of the PCM is changed very little when examined by electron microscopy. In contrast to salt extraction, reduction of functional Dgrip91 has a differential effect upon the loss of centrosomal antigens. CNN remains in distinct bodies at most of the well focused poles, indicating that its centrosomal association is not dependent upon the presence of the gamma-TuRC. The defects of centrosomin(cnn) mutants have been characterized for a number of alleles that show maternal effects and male sterility. These indicate that its function is required for the integrity of both centrosomal and centriolar structures. Syncytial embryos derived from centrosomin mutant mothers undertake up to 12 rounds of mitosis upon spindles whose poles have no astral microtubules and have very little or none of the centrosomal proteins CP60, CP190, or gamma-tubulin. Together, these data imply that CNN appears to be more important in holding the structure of the centrosome together than does the gamma-TuRC, and this is perhaps to be expected from the predicted coiled-coil nature of CNN (Barbosa, 2000).

It is clear that mitotic spindles can form and function in the absence of centrosomes. Repeated rounds of mitosis are known to take place in the absence of centrosomes in the unfertilized eggs of Sciara flies. Moreover, in Drosophila eggs derived from polo mothers, the four products of female meiosis are capable of undergoing many rounds of mitosis on acentriolar spindles. These spindles strongly resemble the meiotic spindles of female Drosophila in which gamma-tubulin cannot be detected by immunostaining at these spindle poles, even though it is apparently needed for spindle function. The ability to build a functional spindle in Xenopus extracts in the absence of centrosomes is also well documented and requires minus end directed motors such as dynein to focus the poles. The consequences of removing centrosomes from cells that have robust checkpoints to monitor spindle assembly can vary, and could reflect either or both the cell line studied and exactly how the experiment was performed. Microsurgical removal of centrosomes has been reported to block future cycles of cell division. However, laser directed ablation of either one or both centrosomes does not prevent assembly of spindles that could successfully undertake anaphase. The high mitotic index resulting from partial disruption of the centrosome in dd4 mutants suggests a mitotic delay likely to result from activation of the spindle integrity checkpoint known to be functional in larval brain cells (Barbosa, 2000).

The distribution of the Asp following the apparent breakdown of the gamma-TuRC gives insight into how these proteins might cooperate in microtubule nucleation. It is known that, following KI depletion of centrosomes, their ability to organize asters of microtubules can only be restored by supplying a complementary cytoplasmic extract that contains both the gamma-TuRC and functional Asp protein. In wild-type cells, Asp forms a hemispherical cup-like structure on the face of the spindle microtubules suggesting that it is contacting the minus ends of these, and not the astral microtubules. Astral microtubules are not seen in dd4 mutant cells at either the light or EM levels, and the spindle poles exhibit varying degrees of disorganization. Nevertheless, the Asp protein is invariably present at the spindle poles even in those extreme cases where individual bundles of microtubules are no longer held together at a single poorly focused pole. In such cases Asp appears at the very tips of these tubules as if it is providing some capping property to their minus ends (Barbosa, 2000).

It is difficult to compare the effects of gamma-tub23C and dd4 mutations upon the structure of the centrosome itself. Nevertheless, although the centrosome had abnormal morphology judged by the distribution of CP190 (Bx63 antigen) in the gamma-tub23C mutant, the antigen was only noted as being at pole-like structures. Unfortunately, there are currently no known mutants of the CP190 gene, and its function remains unknown. CP190 exists in a complex with CP60, and both proteins are known to be nuclear during interphase and move onto centrosomes at mitosis. The extent of interaction between these proteins and gamma-tubulin is also unclear. Two complexes containing gamma-tubulin have been purified from Drosophila embryos; the 3 MD gamma-TuRC itself, and a smaller complex of 240 kD that appears to be a subunit of the larger one. The CP190-CP60 complex does not appear to be present in either of these gamma-tubulin complexes from which it was separable by gel-filtration. However, low levels of gamma-tubulin can be detected in the eluate from immunoaffinity columns constructed from antibodies to CP190 and CP60. This has led to the speculation that although these proteins may not assemble with each other in stoichiometric ratios, they may still show interactions, either on an affinity column in vitro or during centrosome assembly. Consistent with this is the observation that following loss of the majority of the gamma-TuRC from the centrosome in dd41, some CP190 remains in the centrosome, whereas some dissociates and clusters in punctate arrays in the region of the spindle occupied by the condensed chromosomes. In this sense CP190 may be obeying elements of a nuclear localization signal that directs its interphase location, the nuclear envelope undergoing incomplete breakdown during mitosis in Drosophila to form a fenestrated envelope around the spindle (Barbosa, 2000).

Despite the differences in fixation procedures, several aspects of the ultrastructure of the mitotic apparatus in dd4 cells as seen by electron microscopy, such as the microtubule density, are concordant with observations by immunofluorescence. Chromosomes are abnormally condensed and the number and density of spindle microtubules is greatly reduced in the mutant cells. The dispersion of the gamma-tubulin, which is assumed to be the primary consequence of the dd4 mutations, is reflected by disorganization of the PCM and altered centriole morphology. Some gamma-tubulin has been shown to be localized to the core of the centriole, and inactivation of the gamma-tubulin gene in Paramecium leads to inhibition of the duplication of the related structures, the basal bodies. The finding of fewer than four centrioles in the serial EM sections of some dd4 mutant cells suggests a failure of centriole duplication. However, the failure to find centrioles at one of the poles in six dd4 mutant cells suggests that centrosome separation is also dependent upon a functional gamma-TuRC. This may be related to a function in correctly holding centrioles together, because mother and daughter centrioles are rarely perpendicular. The extent to which other centrosomal components found principally in the PCM can contribute to structure of the centriole is not clear. Nonetheless, it is interesting in this context that an isoform of CNN expressed during spermatogenesis is localized both to the centrosomes and to the basal body and has been shown by mutational analysis to be required for the organization of the flagellar axoneme that develops from the spermatid basal body (Barbosa, 2000).

Polo kinase and Asp are needed to promote the mitotic organizing activity of centrosomes

Interfering with the activity of polo-like kinases can lead to the formation of monopolar spindles. Polo-like kinases also regulate mitotic entry, activation of the anaphase-promoting complex and the necessary preconditions for cytokinesis. Similarities between the phenotypes of the Drosophila mutants asp and polo point towards a common role in spindle pole function. The abnormal spindles of asp mutants are bipolar but have disorganized broad poles at which gamma-tubulin has an abnormal distribution. Moreover, the synergism of polo1;aspE3 double mutants indicates a possible involvement of these genes in a common process. Asp is a microtubule-associated protein of relative molecular mass 220,000 (Mr 220K) found at the face of the centrosome that contacts spindle microtubules. In partially purified centrosomes, it is required with gamma-tubulin to organize microtubule asters. Asp is the previously identified Mr 220K substrate of Polo kinase. Polo phosphorylates Asp in vitro, converting it into an MPM2 epitope. Polo and Asp proteins immunoprecipitate together and exist as part of a 25-38S complex. Extracts of polo-derived embryos are unable to restore the ability of salt-stripped centrosomes to nucleate microtubule asters. This can be rescued by addition of phosphorylated Asp or active Polo kinase (do Carmo Avides, 2001).

These findings offer a route to understanding the role of phosphorylation in regulating the behaviour of the centrosomal microtubule-organizing centre. Thus, of the three microtubule-associated Polo substrates in Drosophila embryos, one was identified as beta-tubulin and another is now shown to be Asp. Asp becomes an MPM2 epitope when phosphorylated by Polo in vitro, suggesting that its MPM2 reactivity in wild-type but not asp mutant embryos is likely to be a direct consequence of a Polo-mediated phosphorylation event. The phosphorylation of Asp by Cdk1 and MAP kinases might also have physiological significance and, in fact, a genetic interaction has been reported between asp and rolled, a MAP kinase mutant, in Drosophila. Asp and Polo immunoprecipitate together and seem to be components of a complex that sediments at 25-35 S and has an estimated Mr of 3,000K by gel filtration. The size of the complex suggests that it either contains multimers of Asp and/or Polo or has additional components. It is not know where in the cell this complex might be located but it is known that both Asp and Polo localize to the centrosomes. Polo is likely to be associated with many proteins as it is dispersed both on sucrose gradients and following gel filtration. Polo-like kinases can associate with Cdc25 and the MKLP1/PAV-KLP motor protein. The position at which Cdc25 fractionates by these methods has not been checked, although it is known that PAV-KLP sediments at a position distinct from Asp, suggesting that it is part of a different complex (do Carmo Avides, 2001).

Although Polo, Cdk1 and MAP kinases all phosphorylate Asp in its amino-terminal region, this phosphorylation does not seem to be required for the binding of this segment of Asp to microtubules. Moreover, since Asp is found at spindle poles in polo mutants, it would seem not to require phosphorylation, at least by Polo kinase, for this aspect of its localization. Nevertheless, phosphorylated Asp protein is required for cytoplasmic extracts from polo1-derived embryos to be able to restore microtubule-nucleating activity to preparations of salt-stripped centrosomes. This provides an explanation of why mutations in both polo and asp result in broad spindle poles at which the gamma-tubulin ring complex is not focused, and why polo:asp double mutants have a synergistic interaction. It is proposed that, upon association of Asp and Polo with the centrosome at the onset of prophase, Polo phosphorylates Asp and thereby stimulates its activity to organize microtubules into asters that will later form the spindle poles (do Carmo Avides, 2001).


abnormal spindle: Biological Overview | Evolutionary Homologs | Developmental Biology | Effects of Mutation | References

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.