InteractiveFly: GeneBrief

Monopolar spindle 1: Biological Overview | References

BIOLOGICAL OVERVIEW

Mps1 kinase plays an evolutionary conserved role in the mitotic spindle checkpoint. This system precludes anaphase onset until all chromosomes have successfully attached to spindle microtubules via their kinetochores. Mps1 overexpression in budding yeast is sufficient to trigger a mitotic arrest, which is dependent on the other mitotic checkpoint components, Bub1, Bub3, Mad1, Mad2, and Mad3. Therefore, Mps1 might act at the top of the mitotic checkpoint cascade. Moreover, in contrast to the other mitotic checkpoint components, Mps1 is essential for spindle pole body duplication in budding yeast. Centrosome duplication in mammalian cells might also be controlled by Mps1, but the fission yeast homolog is not required for spindle pole body duplication. Phenotypic characterizations of Mps1 mutant embryos in Drosophila demonstrate that the mitotic spindle checkpoint is defective in these mutants, but do not reveal an involvement in centrosome duplication. In addition, these analyses revealed novel functions. Mps1 is also required for the arrest of cell cycle progression in response to hypoxia. Finally, it was shown that Mps1 and the mitotic spindle checkpoint are responsible for the developmental cell cycle arrest of the three haploid products of female meiosis that are not used as the female pronucleus (Fischer, 2004).

The Drosophila gene CG7643 encodes a Mps1 homolog. The predicted amino acid sequence of Drosophila Mps1 has a higher similarity to vertebrate as compared to yeast orthologs. Similarities are found predominantly in the C-terminal protein kinase domains. Using a piggyBac transposon with an insertion preference clearly distinct from the widely used P elements, an insertion was identified in the Drosophila Mps1 gene. In the following, this allele will be designated as Mps1¹. The piggyBac insertion disrupts the codon of amino acid 26 within the first exon. It is therefore predicted to cause an extensive gene product truncation that removes most of the N-terminal regulatory region and the entire C-terminal kinase domain. In principle, aberrantly spliced readthrough transcripts or transcripts starting within the piggyBac transposon might allow expression of N-terminally truncated Mps1 protein. Quantitative RT-PCR experiments indicated that the abundance of such aberrant transcripts in Mps1¹ mutants, which span the first exon junction downstream of the piggyBac insertion, reached at most 2.5% of the wild-type Mps1 transcript levels. Moreover, deletion analyses with human Mps1 have revealed that N-terminal truncations abolish the normal kinetochore localization (Liu, 2003; Stucke, 2004). Therefore, it is concluded that Mps1¹ is very likely a null allele (Fischer, 2004).

Mps1¹ homozygous progeny of heterozygous parents were found to develop up to the pupal stages. The great majority of the Mps1¹ homozygotes died as pharate adults, but a small fraction eclosed with rough eyes and bent wings. The premature lethality of Mps1¹ homozygotes was completely prevented by a transgene (gEGFP-Mps1) driving expression of an EGFP-Mps1 fusion protein under the control of the Mps1 regulatory region. The recessive lethality associated with Mps1¹ therefore reflects a lack of Mps1 function. Moreover, EGFP-Mps1 is a functional protein (Fischer, 2004).

The relatively late stage of Mps1¹ lethality suggested that Mps1 is not absolutely essential for progression through the cell division cycle. However, phenotypic consequences in Mps1¹ homozygotes might also have been masked by a maternal Mps1⁺ contribution. Therefore, Mps1¹ germline clones were generated, and Mps1¹ mutants lacking both maternal and zygotic function were analyzed. Forty-five percent of these mutants still developed up to the pupal stages, suggesting that cell proliferation can proceed substantially in the absence of Mps1 function. In contrast, embryonic lethality was found to result from a mutation in zebrafish mps1 (Poss, 2002; Fischer, 2004).

Since Mps1¹ mutant cells are able to progress through division cycles, it appeared unlikely that centrosome duplication was affected. To directly analyze centrosome behavior in progeny of Mps1¹ germline clones, embryos were immunolabeled with an antibody against γ-tubulin. These stainings did not reveal centrosome abnormalities. Moreover, a D-TACC-GFP transgene, which expresses a centrosomal protein, was introduced into the Mps1¹ mutant background, and centrosome duplication was analyzed during the early embryonic, syncytial cell cycles by in vivo imaging. Again, centrosome behavior was found to be normal (Fischer, 2004).

While the mouse Mps1 homolog (ESK/mMps1) has been proposed to be important for centrosome duplication, an initial study of human Mps1 (TTK/PYT/hMps1) did not confirm such an involvement (Stucke, 2002). Subsequent work has pointed to a higher hMps1 requirement for the mitotic spindle checkpoint than for normal centrosome behavior (Fisk, 2003). Residual Mps1 function allowing centrosome duplication in the Drosophila Mps1¹ mutant is not ruled out. However, since the expression of Mps1 function is rather difficult to envisage in the light of the curren characterization of the Mps1¹ allele, it is readily conceivable that, like the fission yeast homolog, Drosophila Mps1 is not required for normal spindle pole behavior (Fischer, 2004).

While all analyzed Mps1 homologs localize to kinetochores during prometaphase, conflicting observations have been reported concerning centrosome and nuclear envelope localization in human cells. To evaluate the localization of Drosophila Mps1, gEGFP-Mps1 embryos were analyzed by in vivo imaging and immunolabeling. During interphase in syncytial blastoderm stage embryos, EGFP-Mps1 was excluded from the nucleus and enriched on centrosomes. During prophase, centrosome signals weakened. In parallel, signals appeared on kinetochores, as confirmed by double labeling with an antibody against the centromere protein Cid, the Drosophila Cenp-A homolog. Until the metaphase-to-anaphase transition, EGFP-Mps1 gradually disappeared again from kinetochores and became increasingly enriched on spindles and spindle poles. In anaphase, kinetochore signals were no longer detected. Very similar observations were also made in embryos after cellularization (Fischer, 2004).

Consistent with the observed kinetochore localization during early mitosis, Drosophila Mps1 was found to be required for the mitotic spindle checkpoint, as previously described for all the other analyzed homologs. Inhibition of mitotic spindle formation by colcemid failed to induce a mitotic arrest in the progeny of Mps1¹ germline clones (Fischer, 2004).

Whether Mps1 regulates the dynamics of progression through mitosis in unperturbed conditions was also analyzed. Time-lapse analyses with Mps1¹ germline clone progeny expressing histone H2AvD-GFP demonstrated that the metaphase-to-anaphase transition occurred prematurely during the syncytial cycles. Analyses with cultured vertebrate cells have indicated that elimination of Mad2 and BubR1 (but not of Mad1, Bub1, and Bub3) also results in a premature onset of anaphase and severe chromosome segregation defects in most cells. In the Drosophila Mps1¹ mutants, chromosome segregation was still normal in most cells. Nevertheless, an increased frequency of occasional division failures of individual nuclei was readily detected in syncytial embryos. During these aberrant mitoses, some or even all chromosomes failed to display poleward movements during anaphase, suggesting that they had not yet been properly attached to the spindle by the time of the premature metaphase-to-anaphase transition. These mitotic division failures eventually resulted in the elimination of aberrant nuclei from the superficial nuclear layer -- a characteristic process during the syncytial blastoderm cycles that prevents formation of aneuploid cells during cellularization. Analyses of fixed embryos confirmed that about 45% of the fertilized embryos displayed severely reduced numbers of nuclei with a highly irregular distribution and appearance, preventing syncytial and cellular blastoderm formation. Another 45% of the fertilized embryos appeared to progress successfully beyond cellularization with no or few irregularities in the distribution and appearance of nuclei or mitotic figures. These irregularities were observed predominantly within the polar regions (Fischer, 2004).

Inhibition of the mitotic checkpoint by overexpression of a dominant-negative BubR1 kinase in cultured human cells has previously been shown to advance not only the onset of anaphase but also cyclin B1 degradation so that this mitotic cyclin was degraded at the same time as cyclin A. Immunolabeling of cellularized Mps1¹ germline clone progeny also revealed a simultaneous disappearance of Cyclin A and B. In contrast, in unperturbed human cells, as well as in cellularized Drosophila embryos and other organisms, the mitotic degradation of A-type cyclins occurs before that of the B-type cyclins. Therefore, in addition to BubR1, Mps1 is also required for the delay of B-type cyclin degradation relative to A-types. The difference in the dynamics of Drosophila Cyclin B and B3 degradation did not appear to be affected in the Mps1¹ mutants (Fischer, 2004).

An RNA interference screen for genes required for survival of anoxia in C. elegans has recently led to the identification of the san-1 gene, which encodes a Mad3 homolog involved in the mitotic spindle checkpoint (Pandey, 2007). In addition, the Mad2 homolog mdf-2 was similarly found to be required for the mitotic spindle checkpoint and anoxia survival. Anoxia arrests mitotic cells not only in C. elegans embryos, but also in Drosophila embryos. In Drosophila, precellularization embryos that are confronted with oxygen limitation during prophase arrest cell cycle progression rapidly and reversibly in metaphase. Hypoxic conditions imposed during all other cell cycle stages induce a reversible arrest in interphase, which is accompanied by abnormal chromatin condensation. It is not known whether the metaphase arrest resulting from hypoxia in Drosophila embryos depends on the function of mitotic spindle checkpoint proteins. Moreover, since the C. elegans genome sequence does not contain a Mps1 homolog, it appeared to be of particular interest to evaluate whether Drosophila Mps1 is required for the hypoxia-induced metaphase arrest. Therefore, stage 5 embryos derived from Mps1¹ germline clones were mixed with Mps1⁺ control embryos expressing a histone H2AvD-GFP transgene, and aliquots of the resulting embryo mixture were incubated for 20 min in hypoxic or normoxic conditions, respectively, before fixation and DNA labeling. Confirming previous analyses, hypoxic Mps1⁺ control embryos (GFP-positive) were arrested either in interphase with abnormally condensed chromosomes or in mitosis with clusters of hypercondensed chromosomes within a mitotic spindle. Most importantly, no Mps1⁺ embryos where observed where all nuclei were in anaphase or telophase in a total of 57 analyzed embryos. In contrast, Mps1¹ embryos (GFP-negative) were frequently observed to be in anaphase or telophase (24% of the 161 analyzed embryos), demonstrating that Mps1 is required for the hypoxia-induced metaphase arrest (Fischer, 2004).

While anaphase and telophase figures are clearly recognizable in the hypoxic Mps1¹ mutants, it is emphasized that these mitotic figures are almost always highly aberrant with frequent chromatin bridges. Moreover, chromosome arms, which appear as straight lines during wild-type anaphase, had a wavy appearance in hypoxic Mps1¹ anaphase figures, hinting at reduced spindle pulling forces (Fischer, 2004).

The dramatic difference in the frequencies of embryos during exit from mitosis (anaphase, telophase), which was apparent in the comparison of hypoxic Mps1⁺ and Mps1¹ embryos, was not observed when the normoxic embryos were compared. However, the dramatic difference was also observed in hypoxic mixtures in which the embryos derived from Mps1¹ germline clones were marked with histone H2AvD-GFP instead of the Mps1⁺ control embryos. It is concluded, therefore, that Mps1 is required for the metaphase arrest that results in response to hypoxic conditions imposed during prophase. Mps1 does not appear to play a role in the hypoxia-induced cell cycle arrest during interphase. A comparable high fraction of Mps1⁺ and Mps1¹ embryos arrested in interphase with the characteristic abnormal chromatin appearance was observed in the hypoxic embryo mixture (Fischer, 2004).

In vertebrates, the mitotic spindle checkpoint is known to be involved in the physiological arrest of mature oocytes during metaphase of the second meiotic division. In Drosophila, mature oocytes arrest during metaphase of the first meiotic division. The early developmental arrest during cycle 1 or 2, which was apparent in 9% of the fertilized embryos derived from Mps1¹ female germline clones, might therefore reflect defective regulation of the female meiotic divisions. However, the physiological arrest during metaphase of meiosis I did not appear to be compromised in Mps1¹ mutant oocytes. DNA labeling of mass isolated mature Mps1¹ mutant oocytes. Moreover, after in vitro activation, normal meiotic figures were observed in Mps1¹ mutants. While the possibility that all or some Mps1¹ mutant oocytes are subtly affected cannot be excluded, these observations suggest that Mps1 is required neither for the physiological arrest of mature oocytes during metaphase I nor for completion of the meiotic divisions after release from this metaphase I arrest (Fischer, 2004).

While female meiosis appeared to be largely successful, the subsequent behavior of the three haploid meiotic nuclei, which are segregated away from the female pronucleus, was found to be abnormal in eggs derived from Mps1¹ female germline clones. This abnormal behavior was observed in all of the Mps1¹ progeny, irrespective of whether the zygotic nuclei succeeded at progressing through the syncytial cycles or not. During wild-type development, the innermost one of the four haploid nuclei generated by female meiosis develops into the pronucleus, whereas the peripheral, outer three polar body nuclei become reorganized into a characteristic condensed chromosome bouquet. In the Drosophila egg, where cytokinesis is omitted until after cellularization, these chromosomes are not extruded. Chromosome condensation in the polar body nuclei occurs concomitantly with entry into mitosis 1 after the first round of DNA replication. Condensation of the polar body chromatin is not accompanied by assembly of bipolar mitotic spindles, while the duplicated, sperm-derived centrosomes organize such a spindle around the female and male pronuclei chromosomes. In wild-type embryos, a very strong enrichment is observed of the mitotic spindle checkpoint components EGFP-Mps1, BubR1, Mad2-GFP, and GFP-Fizzy within the pericentromeric regions of the condensed polar body chromosomes. Moreover, anti-pH3 labeling was also strikingly intense within this chromosomal region. Phosphorylation of histone H3 is thought to result from kinase activity of the Survivin-Incenp-Aurora B complex, which is known to be required for kinetochore localization of mitotic spindle checkpoint components. Immunolabeling with antibodies against Aurora B and Incenp indicated that these proteins were also strongly enriched within the pericentromeric region of the bouquet chromosomes. This prominent enrichment of checkpoint components and pH3 on polar body chromosomes perdured in wild-type embryos throughout the early syncytial stages. During these stages, the polar bodies remain condensed, while the zygotic nuclei progress through mitotic cycles. These observations suggested that polar bodies might be kept inactive and arrested in mitosis by the mitotic spindle checkpoint in wild-type embryogenesis. Consistent with this notion, an enrichment of Cyclin B was observed in the central domain of the bouquet. Moreover, this notion predicts that the polar body chromosomes escape from the arrest in Mps1¹ mutants. Indeed, instead of the bouquet-like arrangement of condensed polar body chromosomes that is characteristic for wild-type embryos, one or two giant nuclei were observed in all of the progeny derived from Mps1¹ germline clone females. These giant nuclei contained decondensed chromatin, which was not labeled with anti-BubR1, Mad2-GFP, and anti-pH3 in the great majority of the embryos. It is pointed out that, during the syncytial blastoderm cycles in the zygotic nuclei, the behavior of BubR1 was not affected by the Mps1¹ mutation, in contrast to the findings observed with immunodepleted Xenopus extracts. The absence of mitotic spindle checkpoint components from the abnormal giant nuclei in the Mps1¹ mutants therefore presumably results indirectly from the failure to arrest during mitosis 1. The intensity of the DNA labeling indicated that the giant nuclei had undergone overreplication. Labeling with an antibody against the constitutive centromere protein CID/Cenp-A resulted in an excessive number of signals, and these were generally located in the outer periphery of the decondensed chromatin, while the expected 12 pairs of centromere signals were detected in the center of wild-type bouquets. Based on estimates of centromere signal numbers, at least four rounds of overreplication had occurred in the analyzed Mps1¹ embryos (Fischer, 2004).

This characterization of Mps1 function in Drosophila has uncovered a previously unknown involvement in the hypoxia stress response. These observations extend recent findings obtained in C. elegans, which does not have an obvious Mps1 homolog but requires other mitotic spindle checkpoint components for efficient protection against aneuploidies resulting from progression through mitosis when oxygen is absent. The mitotic spindle checkpoint is therefore likely of general importance for protection against hypoxia-induced genetic damage in metazoans. In humans, such a role of the mitotic spindle checkpoint might further augment its relevance in the context of tumor progression where angiogenesis is crucial for the elimination of the insufficient blood (and oxygen) supply during initial tumor growth (Fischer, 2004).

Moreover, these analyses have revealed a novel, developmental role of Mps1 during early embryogenesis when it is required for the cell cycle arrest of those three haploid products of female meiosis that are segregated away from the female pronucleus. In Drosophila, these discarded meiotic products are not extruded. They arrest during M phase of the first mitotic cycle and are transformed into a cluster of radially arranged condensed chromosomes. It remains to be clarified why the interdigitated microtubules are not focused into the characteristic, centrosome-free, tapered bipolar spindle, which is organized by chromosomes during the preceding female meiotic divisions. However, the failure to form a functional bipolar spindle might be linked to the activation of the mitotic spindle checkpoint, which keeps the bouquet chromosomes from further cell cycle progression. While these findings provide an initial molecular insight, the important mechanisms that differentiate the female pronucleus from the other haploid products of female meiosis are far from being understood (Fischer, 2004).

The meiotic defects of mutants in the Drosophila mps1 gene reveal a critical role of Mps1 in the segregation of achiasmate homologs

The conserved kinase Mps1 is necessary for the proper functioning of the mitotic and meiotic spindle checkpoints (MSCs), which monitor the integrity of the spindle apparatus and prevent cells from progressing into anaphase until chromosomes are properly aligned on the metaphase plate. In Drosophila, a null allele of the gene encoding Mps1 was recently shown to be required for the proper functioning of the MSC, but it did not appear to exhibit a defect in female meiosis. This study demonstrates that the meiotic mutant ald¹ is a hypomorphic allele of the mps1 gene. Both ald¹ and a P-insertion allele of mps1 exhibit defects in female meiotic chromosome segregation. The observed segregational defects are substantially more severe for pairs of achiasmate homologs, which are normally segregated by the achiasmate (or distributive) segregation system, than they are for chiasmate bivalents. Furthermore, cytological analysis of ald¹ mutant oocytes reveals both a failure in the coorientation of achiasmate homologs at metaphase I and a defect in the maintenance of the chiasmate homolog associations that are normally observed at metaphase I. It is concluded that Mps1 plays an important role in Drosophila female meiosis by regulating processes that are especially critical for ensuring the proper segregation of nonexchange chromosomes (Gilliland, 2005).

The altered disjunction (ald¹) mutation was recovered in 1980 from a screen of ethylmethane-sulfonate-treated autosomes for mutants that induced high levels of meiotic nondisjunction (O'Tousa, 1982). ald¹ is a recessive mutation whose only previously noted defect was chromosome missegregation during female meiosis. Although ald¹ has no effect on either the frequency or distribution of recombination, it does induce high levels of nondisjunction at the first meiotic division. Most of the observed nondisjunction reflects the failure of the achiasmate segregation system, which ensures the segregation of those homologs that fail to recombine (Gilliland, 2005).

The effect of the ald¹ mutation on achiasmate X chromosome segregation can be observed with the multiply inverted balancer chromosome FM7. X-chromosome exchange is strongly suppressed or eliminated in females heterozygous for FM7 and a normal-sequence X chromosome, forcing the X chromosomes into the achiasmate pathway. Whereas only very low frequencies of X-chromosome nondisjunction are observed in FM7/X;ald¹/ald⁺ control females, the frequency of X-chromosome nondisjunction in FM7/X;ald¹/ald¹ females is increased more than 20-fold to 17.8%. An even more dramatic increase in X-chromosome nondisjunction is observed in FM7/X;ald¹/Df females (44.4%), which carry only a single copy of the hypomorphic ald¹ mutation opposite a deletion that includes the ald gene. Indeed, the X-nondisjunction frequency observed in FM7/X;ald¹/Df females approaches the frequency of 50% expected if nonexchange X chromosomes simply segregate at random (Gilliland, 2005).

In contrast to the dramatic effects of the ald¹ mutation on achiasmate-X nondisjunction, relatively low levels of X nondisjunction (5.5%) are observed for chiasmate X chromosomes in X/X;ald¹/ald¹ females. Moreover, the majority of even this low level of X-chromosome nondisjunction most likely reflects the failed segregation of those achiasmate-X bivalents that spontaneously occur in some 5%-10% of normal oocytes. On the basis of an analysis of nondisjunctional progeny recovered from X/X;ald¹/ald¹ females, O'Tousa concluded that only 4% of chiasmate X chromosomes nondisjoin in X/X;ald¹/ald¹ females (Gilliland, 2005).

The segregation of the much smaller, and always nonexchange, 4^th chromosome is also impaired in FM7/X;ald¹/ald¹, FM7/X;ald¹/Df, and X/X;ald¹/ald¹ females. As has been observed for other achiasmate segregation-defective mutants, the frequency of 4^th-chromosome nondisjunction is usually less than half the frequency of X nondisjunction, and a substantial fraction of the 4^th-chromosome nondisjunction appears to result from instances in which the two X chromosomes segregate from the two 4^th chromosomes (Gilliland, 2005).

O'Tousa originally mapped ald to cytological region 90C-D. Through deficiency mapping followed by single-nucleotide polymorphism (SNP) mapping this was narrowed to a 23 kb region in 90C1 bounded by two SNPs. A DrosDel project deficiency with a breakpoint within this region fails to complement ald¹; this failure restricts ald¹ to one of four genes, CG14322, CG7523, CG7643, and CG18212. One of those genes, CG7643, encodes the Drosophila homolog of Mps1 (Winey, 2002). Although the parental chromosome on which ald¹ was induced is no longer available, the region containing these four genes was sequenced from both the ald¹ stock and from a stock bearing another mutant (ncd¹) that is located on the same chromosome arm (3R) and which was induced on the same parental 3^rd chromosome as ald¹. In contrast to the genomic reference sequence and the ncd¹ chromosome, only CG18212 and CG7643 had amino acid changes on the ald¹-bearing chromosome. CG18212 cannot be the ald gene because an imprecise P-element excision that ablates ~90% of the protein-coding sequence fully complemented ald¹ with respect to meiotic nondisjunction (Gilliland, 2005).

Four lines of evidence demonstrate that the ald¹ mutation defines the CG7643 transcription unit. First, when compared to the ncd¹ reference chromosome, the CG7643 locus on the ald¹ chromosome contains one missense mutation, R7H. This mutation occurs within the N-terminal regulatory domain, which is less conserved than the C-terminal kinase domain, but at an amino acid site that is conserved at least as far back as D. pseudoobscura; these findings suggest that this residue may be under functional constraint. Second, a P element (P{GS:13084}) that is inserted at nucleotide 15 of the 5' UTR of CG7643 both displays a dramatic defect in achiasmate segregation when heterozygous with a deficiency that uncovers ald⁺ and fails to complement ald¹ with respect to meiotic nondisjunction. Indeed, the levels of X-chromosome nondisjunction observed in FM7/X;ald¹/P{GS:13084} females are comparable to those observed in FM7/X;ald¹/ald¹ females, and FM7/X;P{GS:13084}/Df females display levels of X-chromosome nondisjunction comparable to those observed in FM7/X;ald¹/Df females. Furthermore, a precise excision (as determined by sequencing) of that P element fully rescues nondisjunction in chiasmate Excision/Df females. Third, both ald¹ and the P{GS:13084} insertion mutant exhibit a failure of the MSC that is similar to that of a previously characterized mutation in the CG7643 gene (Fischer, 2004). Finally, a P-element rescue construct carrying the CG7643 transcription unit plus 1.1 kb of upstream sequence rescues the meiotic nondisjunction defects observed in both FM7/X;ald¹/ald¹ and X/X;ald¹/ald¹ females. These four observations indicate that the ald¹ mutation defines the CG7643/mps1 transcription unit (Gilliland, 2005).

To demonstrate that both the ald¹ mutation and the P{GS:13084} insertion exhibit the expected defect in the MSC, larval neuroblasts were treated with the microtubule depolymerizing agent colchicine. This normally activates the checkpoint and prevents progression into anaphase, as evidenced by the separation of sister chromatids. No examples of precocious sister-chromatid separation were found in ald⁺/ald⁺ larvae, whereas 19.8% of the chromosome spreads observed in ald¹ homozygotes and 34.8% of the mitotic figures in P{GS:13084} homozygotes exhibited obvious sister-chromatid separation. It is curious that both ald¹/ald⁺ and P{GS:13084}/ald⁺ heterozygotes exhibited an intermediate defect: 4.7% of spread nuclei exhibited separated sister chromatids in ald¹/ald⁺, and 7.4% of mitotic figures in P{GS:13084}/ald⁺ heterozygotes displayed separated sister chromatids. This partial dominance was unexpected because the meiotic defect of ald¹ is fully recessive, but it suggests that successful arrest after colchicine treatment requires a full dose of Mps1 protein (Gilliland, 2005).

It is also noted that the P{GS:13084} allele has a stronger mitotic-defect phenotype than ald¹ has, despite the fact that it has a weaker effect on female meiosis than does ald¹. The naïve expectation for a missense mutation like ald¹ is that a normal amount of transcript and protein will be produced but that the resulting protein will be defective. In contrast, a P-element insertion in the 5′ UTR of a locus is more likely to alter the amount of transcript produced, but the protein produced from that transcript is likely to be normal. Therefore, the observation of stronger mitotic defects in P{GS:13084} may suggest that, with respect to Mps1 concentration, mitosis is a more dosage-sensitive process than is meiosis. Such a suggestion is buttressed by the fact that both the P insertion and ald¹ are semidominant with respect to their effects on mitosis but fully recessive in meiosis. Alternatively, it is also possible that the ald¹ mutation alters a protein portion that is more critical for meiosis than it is for mitosis. Finally, the possibility that differences in such quantitative phenotypes as the absolute level of meiotic nondisjunction or the fraction of neuroblasts with separated sister chromatids may be affected by genetic background cannot be ruled out. The analysis of additional alleles will be required to resolve this issue (Gilliland, 2005).

The observation of a defect in the MSC raised the possibility that the meiotic defects observed in ald¹ homozygotes might also be manifested as defects in the control of homolog coorientation or alignment at metaphase I or perhaps even in the control of meiotic progression. To address this issue, meiotic spindle assembly and chromosome alignment were analyzed in ald oocytes. Although no defects were observed in spindle assembly or progression to metaphase I, two types of defects were observed at high frequency among metaphase figures in oocytes from FM7/X;ald¹/ald¹ and FM7/X;ald¹/Df females (Gilliland, 2005).

First, among those oocytes that appeared arrested at metaphase, figures were often observed in which both X or 4^th chromosomes were oriented toward the same pole. For example, among the 14 metaphase figures observed in FM7/X;ald¹/Df oocytes, six were nondisjunctional (in four cases, both X chromosomes were proceeding to the same pole, and the 4^th chromosomes were segregating normally; in one case, both 4^th chromosomes were segregating to the same pole, and the X chromosomes were segregating normally; and in one case, both X chromosomes were segregating to one pole, and the two 4^th chromosomes were segregating to the other pole). These observations allow a cytological estimate of the frequency of meiosis I nondisjunction in FM7/X;ald¹/Df females (36% X- and 14% 4^th-chromosome nondisjunction); this estimate is in good agreement with the genetic estimates of X- and 4^th-chromosome nondisjunction (44% X and 26% 4^th). Such maloriented figures were also observed for FM7/X;ald¹/ald¹ oocytes, and the nondisjunction frequencies estimated by cytological analysis (14% X- and 7% 4^th-chromosome nondisjunction) correlate well with the frequencies of nondisjunction obtained by genetic analysis (18% X- and 9% 4^th-chromosome nondisjunction). These data demonstrate that the achiasmate nondisjunction observed in ald hemizygotes or homozygotes is the result of a failure of achiasmate homologous coorientation at metaphase I (Gilliland, 2005).

Second, despite normal levels of meiotic recombination, chiasmate autosomes of oocytes from FM7/X;ald¹/ald¹ and FM7/X;ald¹/Df females often display a precocious separation that is normally not observed in metaphase oocytes. In the metaphase of normal oocytes, the chiasmate autosomes are usually observed as a single dense chromosomal mass at the metaphase I plate; however, in rare cases, an obvious chiasmate connection can be observed to hold the two masses, corresponding to autosomal bivalents, together at the metaphase plate. The abnormal figures observed in FM7/X;ald¹/ald¹ and FM7/X;ald¹/Df oocytes fall into two classes: one in which the two pairs of autosomes are fully separated but connected by wispy threads of chromatin and another in which a complete separation of the autosomes creates what appear to be anaphase (or anaphase-like) figures. Although such figures were occasionally observed in FM7/X ald⁺/ald⁺ oocytes (7% frequency), they are observed at much higher frequencies in FM7/X;ald¹/ald¹ and FM7/X;ald¹/Df oocytes (~30% for both genotypes). Taken together, these data suggest two prominent defects in ald/mps1 mutant oocytes: one in the alignment of achiasmate chromosome pairs at the metaphase plate and the other in the maintenance of chiasmate connections between the autosomal arms. Moreover, although there has been some debate in the literature as to the role of mammalian mps1 in centrosome regulation, this study has clearly demonstrated a defect of mps1 in an acentriolar cell division, namely meiosis I. This finding is consistent with the failure of Fischer (2004) to find evidence of a mitotic defect in centrosome regulation in mps1 null flies (Gilliland, 2005).

It is noted that Fischer observed neither of these defects in their analysis of the meiotic effects of a lethal and presumably null allele of mps1 in germline clones. There are at least two explanations for the differences in the findings. First, this analysis focused on FM7/X oocytes in which X-chromosome exchange has been suppressed. It is possible that such defects might be ameliorated or absent in females in which X-chromosome exchange occurred at normal frequencies, as was the case in the Fischer study. Indeed, in this own cytological analysis, out of 17 metaphase X/X;ald¹/ald¹ oocyte nuclei, only one example of complete homolog separation, which generated an anaphase I (or anaphase I-like) figure, was observed. Five cases were observed of precocious homolog separation, but because the phenotype was more subtle than in the achiasmate case, this defect might have gone unnoticed but for the authors' experience in the analysis of FM7/X oocytes. (No cases of precocious homolog separation were observed in 50 metaphase figures obtained from X/X;ald⁺/ald⁺ oocytes.) It thus seems likely that the cytological defects observed are most easily visualized in a sensitized meiotic system in which the X chromosomes are rendered achiasmate in all oocytes. Alternatively, it may well be that the control of meiotic progression in the presence of a reduced level of Mps1 function, as would be the case in oocytes homozygous for the hypomorphic ald¹ allele, has different effects than those of ablating the protein entirely, or that the mps1 allele examined by Fischer primarily affects mitotic function without impairing meiotic function (Gilliland, 2005).

These observations parallel the findings of an elevated level of meiotic nondisjunction in zebrafish homozygotes for a temperature-sensitive hypomorphic allele of mps1 (Poss, 2004) and suggest that the germ-cell aneuploidy observed by those workers was in fact the result of defects in chromosome segregation at the first meiotic division. They also confirm the observations of Straight (2000), who described defects in meiotic chromosome segregation in mps1 mutant yeast. Moreover, all of these studies lend further significance to the yeast studies of Schonn, who observed increased levels of meiotic nondisjunction in yeast deficient in the MSC protein Mad2 (Schonn, 2000; Schonn, 2003). Indeed, several lines of evidence have demonstrated that many of the proteins required for cell-cycle control also participate in the regulation of meiosis. These data suggest that MSC proteins play a critical role in assessing proper centromere coorientation and in maintaining homologous-chromosome associations at metaphase I. It is further proposed that such functions may be especially critical in terms of mediating the segregation of achiasmate chromosomes, whose proper orientation is not mediated by chiasmata (Gilliland, 2005).

The multiple roles of mps1 in Drosophila female meiosis

The Drosophila gene ald encodes the fly ortholog of mps1, a conserved kinetochore-associated protein kinase required for the meiotic and mitotic spindle assembly checkpoints. Using live imaging, this study demonstrated that oocytes lacking Ald/Mps1 (hereafter referred to as Ald) protein enter anaphase I immediately upon completing spindle formation, in a fashion that does not allow sufficient time for nonexchange homologs to complete their normal partitioning to opposite half spindles. This observation can explain the heightened sensitivity of nonexchange chromosomes to the meiotic effects of hypomorphic ald alleles. In one of the first studies of the female meiotic kinetochore, this study shows that Ald localizes to the outer edge of meiotic kinetochores after germinal vesicle breakdown, where it is often observed to be extended well away from the chromosomes. Ald also localizes to numerous filaments throughout the oocyte. These filaments, which are not observed in mitotic cells, also contain the outer kinetochore protein kinase Polo, but not the inner kinetochore proteins Incenp or Aurora-B. These filaments polymerize during early germinal vesicle breakdown, perhaps as a means of storing excess outer kinetochore kinases during early embryonic development (Gilliland, 2007. Full text of article).

The progression of meiosis is a tightly temporally regulated event and the study of cell cycle regulation is greatly enhanced by being able to visualize meiosis in living cells. In this study, a defect in a kinetochore component results in the precocious entry into anaphase, presumably as a result of the early activation of Separase, as well as the mis-segregation of nonexchange chromosomes. The greater sensitivity of nonexchange chromosomes to ald mutants is explained by noting that while chiasmate chromosomes appear to co-orient immediately upon germinal vesicle breakdown (GVBD), nonexchange chromosomes require more time to achieve proper biorientation. The test of this hypothesis lies in the relative timing of these two events, data which are straightforwardly obtained from live imaging but difficult, and considerably more ambiguous, to obtain only from the examination of fixed images (Gilliland, 2007).

It was also observed that ald mutant oocytes can enter an anaphase-like configuration practically as soon as the spindle is formed, so it is believed that this requirement is satisfied. This interpretation is strengthened by the three genotypes examined by live imaging. For the two genotypes with the highest rates of nondisjunction (NDJ) (ald¹/Df(3R)AN6 and ald^Exc23/Df(3R)AN6), every oocyte examined appeared to undergo premature separation, while the weakest genotype (ald¹/ald¹) only lost cohesion in 43% of oocytes. This is consistent with the NDJ data, as homologs segregating at random are expected to nondisjoin half of the time, and so if the failure to maintain sister chromatid cohesion is the cause of the NDJ, it should occur at approximately twice the rate of NDJ. While the duration of prometaphase did not correlate with the propensity to nondisjoin, this may reflect a difference between the alleles. The ald¹ oocytes contain only hypomorphic protein, while the ald^Exc23 oocytes still contain a small amount of wild-type protein (Gilliland, 2007).

It is furthermore noted that strong ald alleles also induce chiasmate NDJ at high levels, a fact underscored by the mutant screen that recovered ald^C3. This germline clone screen only recovered mutants that could cause a chiasmate autosome to nondisjoin, and ald was the most commonly hit gene, with four out of 12 recovered mutants containing new alleles of ald. The mechanism cannot be the same as that for nonexchange chromosomes, as chiasmate chromosomes appear to be properly co-oriented after GVBD, without the back-and-forth movement observed for FM7/X. One possibility is that the loss of sister chromatid cohesion when the spindle assembly checkpoint (SAC) is impaired causes all chiasmata to resolve during early prometaphase. This would result in a situation analogous to mutants such as meiW68, the Drosophila homolog of spo11, where recombination has been completely abolished, that also cause high levels of NDJ. Consistent with this interpretation, metaphase I arrest is also bypassed in oocytes with mutants that abolish crossing over as well as chromosomal configurations where chiasmata form but do not establish bipolar tension (Gilliland, 2007).

This study has also extended the previously known mitotic kinetochore association of the Ald protein to female meiosis. It was surprising that in some, but not all, cases, the kinetochore staining of Ald clearly extended well outside the centromere region identified by Cid. One possibility is that the very rapid cycling of Ald through the kinetochore (Howell, 2004) proceeds by transport of the protein along attached microtubules. If this is the case, the extended kinetochores highlighted by Ald staining represent protein that has been released from the kinetochore but has not yet detached from the microtubule. Alternatively, the finding of such kinetochores evokes the model of stretched centromeric DNA observed during mitosis using the LacO/LacI-GFP system in budding yeast. For example, LacO arrays inserted closer than 9 kb away from the centromere were found to transiently separate from their homolog by up to 0.8 microm during metaphase, while arrays inserted 13 kb or further away from the centromere were separated by only 0.3 microm. This finding was interpreted as indicating the centromere-proximal chromatin was being stretched, possibly as part of how the kinetochore senses tension. In light of this finding, an alternative interpretation of the extended Ald staining result is that tension on a kinetochore is able to change its shape, and therefore the cytology may reflect the SAC being in an on or off configuration. As checkpoint signaling requires the kinase activity of Mps1 (Abrieu, 2001), finding this protein along the extended kinetochore structure means it is well positioned to perform this role (Gilliland, 2007).

However, it is noted that neither DAPI fluorescence nor Cid localization to these extended kinetochores was observed, and it cannot be proven that there is DNA present in these structures. It may be that the structure highlighted by extended Ald staining was composed solely of proteins, and represents a kinetochore that is structurally different from the mitotic kinetochore. The Drosophila female meiotic kinetochore has not previously been examined by electron microscopy, and given the lack of centrosomes, the female meiotic kinetochore may have important structural differences as well. If this is the case, then the Drosophila ovary should prove well suited to the study of this kinetochore. In addition to the possibility of monitoring the relative positioning of different proteins on the spindle via immunofluorescence, the well-established use of germline clones would allow for studying the effects on meiosis of mutations in genes that would be lethal during development (Gilliland, 2007).

It is also noted that the localization of Aurora-B and Incenp is different from what was previously reported in mitosis. In mitosis, Aurora-B is a chromosomal passenger protein that is essential to the function of the SAC, and migrates from the kinetochore to the spindle midzone at the metaphase/anaphase transition. This study showed that in female meiosis, this migration happens early in prometaphase, prior to checkpoint inactivation, and localizes to either the midzone (for Incenp) or to a stripe along only one side of the spindle (for Aurora-B). That these two proteins form a different final configuration shows they are no longer present in the same complex at metaphase arrest. It is also noted that the localization is slightly different from that previously reported by Jang (2005) in Drosophila oocytes. In that study, they were unable to find localization of Aurora-B and Incenp to the kinetochore during prometaphase, and they did not see a stripe of Aurora-B staining. This study found kinetochore localization only during the very earliest stages of prometaphase, before the 4 chromosomes have begun moving out from the main chromosomal mass, which was an earlier stage than the images presented in their paper. Also, this study used achiasmate FM7/X oocytes, while Jang used chiasmate X/X oocytes. That the Aurora-B stripe at metaphase arrest appeared along the side of the oocyte that the nonexchange X chromosomes were associated with suggests that this is a feature of achiasmate chromosomes, and may be why Jang did not observe it. It also suggests that Aurora-B may play a role in balancing nonexchange homologs on the meiotic spindle (Gilliland, 2007).

Finally, the identification of Ald-containing filaments represents a novel structure in the Drosophila oocyte, as well as a novel functionality for both Ald and Polo kinases. That these filaments could be found in fixed and unfixed polo-GFP oocytes, as well as fixed wild-type oocytes, indicates that they are neither an artifact of fixation nor of the GFP construct. While both Polo and Ald are kinetochore-associated proteins, of the other kinetochore-associated proteins examined (Cid, Aurora-B, Incenp, Bub1, BubR1, Bub3, and Mad2), only BubR1 localized to a minority of filaments (Gilliland, 2007).

The filaments appear to be sequestering Ald protein, but do not require Ald for their formation. This is clearly indicated by the Polo-GFP/+; ald^Exc23 oocytes, in which filaments were barely detectable by Ald staining, but appeared unaffected when visualized by Polo-GFP staining. Therefore, it is concluded that the reduction of filaments in the various mutant backgrounds must be due to a failure to detect the filaments, even though the filaments are still present. The occasional filament seen in some images confirms this interpretation, as only those filaments that incorporated enough Ald protein over a small region would be classified as a linear structure. Doing the reciprocal experiment of staining for Ald in the presence of a Polo mutant cannot be done due to the sterility of those mutants, but may be possible to conduct through the use of germline clones. Interestingly, the ald¹ hemizygotes had fewer filaments than ald⁺ hemizygotes despite having similar levels of protein. One possibility is that the ald¹ allele cannot be sequestered as efficiently as ald⁺. Since Mps1 is phosphorylated during checkpoint activation (Palframan, 2006), one possibility is that the ald¹ mutation disrupts a phosphorylation site. Indeed, there is an in silico-predicted phosphorylation site that is disrupted in the ald¹ mutant for the threonine-9 codon, suggesting that phosphorylation may be involved in regulating the incorporation of Ald into these filaments (Gilliland, 2007).

How these filaments are formed is an interesting problem. It is proposed that the polymerization is triggered by a signal propagating through the oocyte, and that this signal occurs during germinal vesicle breakdown. However, what that signal is, and what the targets of that signal are, remain to be determined. It is also unclear whether the formation of filaments is necessary for the initiation of GVBD, or a downstream consequence of it. The presence of BubR1 foci at the tips of many of the filaments as they are forming also suggests that other proteins are involved in their construction; it is noted that BubR1 was the only protein (besides Ald and Polo) to incorporate into the filaments at all. A further question, as these filaments appear to disassemble by meiosis II, is what signal initiates the disassembly, and what proteins are required to respond to that signal. Finally, the function of these filaments is unknown. While there is no direct evidence for the function of these filaments, several possibilities suggest themselves. One possibility is that, because many SAC proteins (including Mps1 and Polo) are degraded as part of checkpoint inactivation, the filaments may protect the maternal load of these proteins during this stage. However, this interpretation is not consistent with the filaments dispersing by meiosis II, since the proteins would still be susceptible to degradation during the syncytial nuclear divisions. Another possibility is that the filaments allow the cell to regulate the activity of these proteins during meiosis I. One approach to answer these questions would use immunoprecipitation to isolate these filaments and determine what other proteins are present, followed by examination of mutant alleles of those component proteins. The mechanism that these filaments carry out could likely be inferred by the meiotic defects of mutants in other genes involved in filament formation (Gilliland, 2007).

Spatiotemporal control of mitosis by the conserved spindle matrix protein Megator

A putative spindle matrix has been hypothesized to mediate chromosome motion, but its existence and functionality remain controversial. This report shows that Megator (Mtor), the Drosophila melanogaster counterpart of the human nuclear pore complex protein translocated promoter region (Tpr), and the spindle assembly checkpoint (SAC) protein Mad2 form a conserved complex that localizes to a nuclear derived spindle matrix in living cells. Fluorescence recovery after photobleaching experiments supports that Mtor is retained around spindle microtubules, where it shows distinct dynamic properties. Mtor/Tpr promotes the recruitment of Mad2 and Mps1 but not Mad1 to unattached kinetochores (KTs), mediating normal mitotic duration and SAC response. At anaphase, Mtor plays a role in spindle elongation, thereby affecting normal chromosome movement. It is proposed that Mtor/Tpr functions as a spatial regulator of the SAC ensuring the efficient recruitment of Mad2 to unattached KTs at the onset of mitosis and proper spindle maturation, whereas enrichment of Mad2 in a spindle matrix helps confine the action of a diffusible 'wait anaphase' signal to the vicinity of the spindle. It is also suggested that the regulatory role of Mtor upon Mad2 is indirect and may be catalyzed by Mps1 (Lince-Faria, 2009).

The mitotic spindle is composed of dynamic microtubules (MTs) and associated proteins that mediate chromosome segregation during mitosis. The requirement of an additional stationary or elastic structure forming a spindle matrix where molecular motors slide MTs has long been proposed to power chromosome motion and account for incompletely understood features of mitotic spindle dynamics. However, definitive evidence for its existence in living cells or on its biochemical nature and whether it plays a direct role during mitosis has been missing (Lince-Faria, 2009).

A functional spindle matrix would be expected to (a) form a fusiform structure coalescent with spindle MTs, (b) persist in the absence of MTs, (c) be resilient in response to changes of spindle shape and length, and (d) affect spindle assembly and/or function if one or more of its components are perturbed. In Drosophila, a complex of at least four nuclear proteins, Skeletor, Megator (Mtor), Chromator, and EAST (enhanced adult sensory threshold), form a putative spindle matrix that persists in the absence of MTs in fixed preparations). From this complex, Mtor is the only protein that shows clear sequence conservation with proteins in other organisms, such as the nuclear pore complex (NPC) protein translocated promoter region (Tpr) in mammals, its respective counterparts Mlp1 and Mlp2 in yeast, and nuclear pore anchor in plants. NPC proteins, including Mtor/Tpr orthologues in yeast, have been shown to functionally interact with spindle assembly checkpoint (SAC) components. The SAC ensures correct chromosome segregation by providing time for proper kinetochore (KT) attachments to spindle MTs while inhibiting the activity of the anaphase-promoting complex/cyclosome (Lince-Faria, 2009).

Assuming that any critical function by the spindle matrix is widely conserved, this study focused on understanding the mitotic role of Mtor in living Drosophila somatic cells. The results provide a new conceptual view of a spindle matrix not as a rigid structural scaffold but as a spatial determinant of key mitotic regulators (Lince-Faria, 2009).

Mtor localizes to a dynamic nuclear derived spindle matrix in living cells. To investigate the localization of Mtor in living cells, a Drosophila S2 cell line was generated stably coexpressing Mtor-mCherry and GFP-α-tubulin. Mtor-mCherry is nuclear in interphase and at nuclear envelope breakdown (NEB) reorganizes into a fusiform structure coalescent with spindle MTs. Mtor-mCherry shows a highly adaptable morphology in response to changes in spindle shape and dynamics throughout mitosis, which is inconsistent with a static structure. Similar to endogenous Mtor, Mtor-mCherry retracts and loses the fusiform shape upon MT depolymerization but is retained in a conspicuous milieu around chromosomes, suggesting that MTs exert a pushing force on the Mtor-defined matrix (Lince-Faria, 2009).

Previous electron microscopy analysis revealed the existence of a membranous network surrounding the spindle from prophase to metaphase in S2 cells. This study used immunofluorescence to show that lamin B is not fully disintegrated at this stage. Similar results have recently been reported in living Drosophila embryos and neuroblasts, where a spindle envelope was proposed to limit the diffusion of nuclear derived Nup107 before anaphase. To test whether this membranous network works as a diffusion barrier around the spindle, the dynamic behavior of Mtor-mCherry relative was compared to GFP-α-tubulin and a known MT-associated protein, Jupiter, upon colchicine addition. GFP-α-tubulin or Jupiter-GFP fluorescence is gradually lost from the spindle region with an equivalent gain in the cytoplasm. In contrast, Mtor-mCherry remains confined to the spindle region with no detectable fluorescence gain in the cytoplasm. These results argue against the existence of a diffusion barrier around the metaphase spindle in Drosophila S2 cells and suggest that Mtor is being selectively retained in this region (Lince-Faria, 2009).

FRAP was used to shed light on the dynamic properties of Mtor. In interphase nuclei, there is ~50% recovery of fluorescence in the bleached region with an equivalent loss from a similar unbleached region and undetectable cytoplasmic exchange, suggesting that Mtor in the nucleoplasm is mobile. In mitosis, FRAP of Mtor-mCherry in one half-spindle is mirrored by an equivalent loss of fluorescence from the unbleached half-spindle as if Mtor exchanges between half-spindles. However, this recovery was slower than in interphase nuclei and had a minor contribution from a cytoplasmic pool. In both interphase and mitosis, the recovery curves of Mtor-mCherry fitted a single exponential, suggesting affinity to a yet unidentified substrate, whereas GFP-α-tubulin in the spindle displayed biphasic recovery kinetics and best fit the sum of two exponentials as result of a rapid diffusion phase followed by a slower recovery phase associated with MT turnover. Finally, in S2 cells that sporadically form two spindles in the same cytoplasm, fluorescence exchange was found within the same spindle and from surrounding cytoplasm with no apparent loss from the neighboring unbleached spindle, supporting that Mtor is unable to exchange between two spindles located <10 µm apart. Collectively, these data indicate that Mtor is part of a dynamic, nuclear derived spindle matrix surrounded by a fenestrated membranous system containing lamin B and shows mobility properties that are distinct from MTs and associated proteins (Lince-Faria, 2009).

To address the mitotic role of Mtor, RNAi was used in Drosophila S2 cells stably coexpressing GFP-α-tubulin and the KT marker mCherry-centromere identifier (CID). Mtor-depleted cells show no major spindle defects but typically form a poorly defined metaphase plate as the result of progressing ~15% faster through mitosis when compared with controls. Such problems in completing chromosome congression are corrected if anaphase onset is delayed by treating cells with the proteasome inhibitor MG132. As in Mtor RNAi, S2 cells depleted of the SAC protein Mad2 undergo a faster mitosis. Moreover, Mtor-depleted cells show a lower mitotic index as well as a weakened response to MT depolymerization, suggesting that Mtor is required for proper SAC response (Lince-Faria, 2009).

Quantitative analysis of anaphase revealed a significant attenuation in the velocity of chromosome separation in Mtor-depleted cells by affecting spindle elongation. These results could be accounted for if Mtor is part of a structural scaffold where motor proteins assemble to generate force. However, an alternative hypothesis is that Mtor may function to provide the necessary time for proper maturation of a competent spindle. To test this, anaphase onset was delayed by treating Mtor-depleted cells with MG132, and half-spindle elongation velocity was measured after drug washout. No difference was found in half-spindle elongation velocity between Mtor RNAi and control cells treated with MG132. Additionally, half-spindle elongation in Mad2-depleted cells, which progresses faster through mitosis, was similar to Mtor-depleted cells, supporting the spindle maturation hypothesis (Lince-Faria, 2009).

To shed light on the role of Mtor in SAC response, the recruitment of Mad2 and BubR1 to unattached KTs after Mtor depletion was analyzed. It was found that although BubR1 was unaltered after Mtor depletion, Mad2 KT accumulation was significantly reduced. Decreased Mad2 levels at KTs explain why Mtor-depleted cells enter anaphase prematurely, presumably because it requires binding of fewer MTs to remove all Mad2 from KTs and satisfy the SAC, whereas residual Mad2 at KTs may be sufficient to produce a weakened response to colchicine (Lince-Faria, 2009).

Next, how Mtor regulates the recruitment of Mad2 to KTs in living S2 cells stably coexpressing GFP-α-tubulin and monomeric RFP (mRFP)-Mad2 was investigated. In interphase, mRFP-Mad2 is nuclear, accumulating at unattached KTs and spindles as cells transit into mitosis. The spindle accumulation of Mad2 is thought to result from dynein-dependent poleward transport as MTs attach to KTs. Interestingly, however, it was found that a distinct pool of mRFP-Mad2 localizes to a nuclear derived spindle matrix even when MTs have just started invading the nuclear space. A similar behavior has been observed in vertebrate cells, where GFP-Mad2 accumulates as an ill-defined nuclear derived matrix during early prometaphase after its initial recruitment to unattached KTs. Like Mtor, the retention of Mad2 in the spindle matrix is resistant to MT depolymerization, suggesting that spindle-associated Mad2 is not freely diffusible. mRFP-Mad2 remains associated with the spindle matrix in the absence of Mtor, but it is unable to accumulate at KTs even after MT depolymerization with colchicine. Stable expression of Mtor-mCherry (which is RNAi insensitive) rescues normal Mad2 localization at KTs after Mtor RNAi, indicating that the observed phenotype is specific and supporting that Mtor-mCherry is a functional protein. Lastly, Mtor depletion does not affect normal Mad2 expression levels and vice versa, which rules out unspecific effects of Mtor over Mad2 mRNA transport to the cytoplasm (Lince-Faria, 2009).

The colocalization of Mtor and Mad2 in the spindle matrix suggests that these proteins may interact. Indeed, Mad2 was found to coimmunoprecipitate with Mtor in lysates obtained from Drosophila embryos harvested between 0-3 h after egg laying. Given that Mtor does not specifically accumulate at KTs, this interaction might represent an important regulatory step for the subsequent recruitment of Mad2 to unattached KTs. Several proteins such as Mad1, Rod, Ndc80, or Mps1 are involved in recruiting Mad2 to unattached KTs. Although Mad1, Rod, and Ndc80 are effectively targeted to unattached KTs after Mtor depletion, Mps1 accumulation is significantly reduced. Mps1 kinase activity has been shown to be required to specifically target Mad2 but not Mad1 to unattached KTs in human cells. To investigate whether the same regulatory role upon Mad2 is true in Drosophila, an mps1 kinase-dead (mps1KD) allele was generated by homologous recombination in flies. In agreement with the results in human cells, neuroblasts from mps1KD third instar larvae show reduced or undetectable Mad2 accumulation at KTs upon colchicine treatment. Collectively, these results support that the regulatory role of Mtor upon Mad2 is indirect and may be catalyzed by Mps1 (Lince-Faria, 2009).

Like Tpr, Mad1, Mad2, and Mps1 localize at the NPC during interphase in human cells. During mitosis, Tpr remains associated with the nuclear envelope until prometaphase. Moreover, a fraction of Tpr is associated with the mitotic spindle from late prometaphase until anaphase and is recruited to the reforming nuclear envelope during telophase. This confirms the previous identification of Tpr in isolated human mitotic spindles, no enriched fraction of Tpr was dected that resists MT depolymerization with nocodazole, including KTs. To see whether Tpr has a conserved regulatory role in the recruitment of Mad2 to unattached KTs in human cells, RNAi was used to deplete Tpr in HeLa cells. Like in S2 cells, Tpr RNAi leads to reduced accumulation of Mad2 but not Mad1 to unattached KTs accompanied by a decrease in the normal mitotic index and a weakened SAC response in the presence of nocodazole. Tpr knockdown does not enrich for cells in G2 and slightly increases the number of cells in G1, supporting that the lower mitotic index is not caused by the inability of cells to enter mitosis but rather reflects a faster exit. Moreover, Tpr, Mad1, Mad2, and Mps1 coimmunoprecipitate in mitotic enriched HeLa cell extracts prepared in the presence of nocodazole, extending the results obtained in Drosophila and reinforcing that this complex forms independently of MTs and an intact nuclear envelope. While this paper was under revision, Tpr was independently found to interact with Mad1 and Mad2 in human cells. In agreement with these results, it was proposed that Tpr is important for controlling the SAC but reject the possibility that Tpr is playing a role in mitotic timing. However, quantification of the NEB to anaphase duration in Tpr-depleted cells does show a 25% acceleration of mitosis during this period. Finally, the results are not consistent with a model in which KT-associated Tpr serves as a docking place for Mad1 because Tpr (or Mtor) were not detected at KTs, including those that were positive for Mad1, and no impairment was found in Mad1 KT recruitment in Tpr- or Mtor-depleted cells (Lince-Faria, 2009).

Overall, the results support a model in which Mtor/Tpr acts as a spatial regulator of SAC, ensuring a timely and effective recruitment of Mad2 and Mps1 to unattached KTs as cells enter mitosis. In budding yeast, Mps1 phosphorylates Mad1, which is continuously recycled to KTs from Mlps at NPCs, but N-terminal deletion mutants of Mad1 lacking the Mlp-binding domain have a functional SAC. In humans and Drosophila, Mps1 regulates Mad2 but not Mad1 accumulation at KTs. Because Mad1 localization at KTs does not depend on Mlps/Mtor/Tpr and Mps1 kinase activity, the residual Mad2 at KTs after Mtor/Tpr RNAi possibly corresponds to the Mad1-bound fraction. One possibility is that Mps1 phosphorylation of Mad1 regulates the recruitment of a fast-exchanging pool of Mad2 to KTs. Parallelly, Mtor/Tpr may spatially regulate Mps1 autophosphorylation, which is important for its normal KT accumulation, together with Mad2. The presence of Mad2 in the complex may work as a positive feedback mechanism to ensure continuous Mps1 kinase activity upon SAC activation (Lince-Faria, 2009).

SAC proteins evolved from systems with a closed mitosis like budding yeast, where the spindle assembles inside an intact nuclear envelope into more complex systems like animals and plants, where the nuclear envelope is thought to fully or partially disintegrate during spindle formation, justifying the requirement of a nuclear derived spindle matrix for an effective SAC response. What retains matrix components around the spindle in systems where mitosis is thought to be open remains an intriguing question. In this regard, lamin B was proposed to tether several factors that mediate spindle assembly in Xenopus laevis egg extracts and possibly in human cells. Additionally, a continuous endoplasmic reticulum surrounding the mitotic spindle is thought to be recycled from the nuclear envelope after its disassembly and has been observed in many systems undergoing an open mitosis, including humans. Although such fenestrated membranous systems cannot work as diffusion barriers, it is possible that they indirectly help to generate local gradients or concentrate matrix-affine substrates. The enrichment of Mad2 in the spindle matrix provides an explanation for an unsolved SAC paradigm in which the 'wait anaphase' signal emanating from unattached KTs must be diffusible to prevent premature anaphase onset of already bioriented chromosomes but at the same time is known to be restricted to the vicinity of the spindle (Lince-Faria, 2009).

The proposed role of Mtor/Tpr further supports the necessity of spindle maturation for proper KT-MT attachments and anaphase spindle elongation in which the spindle matrix may help extend the duration of mitosis for the assembly of a competent chromosome segregation machinery. Mtor/Tpr-depleted cells have a weakened SAC response that, as opposed to complete checkpoint loss, may be compatible with cell viability and lead to cancer. The involvement of Tpr in the activation of several oncogenes may translate into an unfavorable combination that facilitates transformation and tumorigenesis in humans (Lince-Faria, 2009).

Silva, R. D., Mirkovic, M., Guilgur, L. G., Rathore, O. S., Martinho, R. G. and Oliveira, R. A. (2018). Absence of the spindle assembly checkpoint restores mitotic fidelity upon loss of sister chromatid cohesion. Curr Biol 28(17):2837-2844. PubMed ID: 30122528

Absence of the spindle assembly checkpoint restores mitotic fidelity upon loss of sister chromatid cohesion

The fidelity of mitosis depends on cohesive forces that keep sister chromatids together. This is mediated by cohesin. Cleavage of cohesin marks anaphase onset. Unscheduled loss of sister chromatid cohesion is prevented by a safeguard mechanism known as the spindle assembly checkpoint (SAC). To identify specific conditions capable of restoring defects associated with cohesion loss, a screen was carried out for genes whose depletion modulates Drosophila wing development when sister chromatid cohesion is impaired. Cohesion deficiency was induced by knockdown of the acetyltransferase separation anxiety (San)/Naa50, a cohesin complex stabilizer. Several genes whose function impacts wing development upon cohesion loss were identified. Surprisingly, knockdown of key SAC proteins, Mad2 and Mps1, suppressed developmental defects associated with San depletion. SAC impairment upon cohesin removal, triggered by San depletion or artificial removal of the cohesin complex, prevented extensive genome shuffling, reduced segregation defects, and restored cell survival. This counterintuitive phenotypic suppression was caused by an intrinsic bias for efficient chromosome biorientation at mitotic entry, coupled with slow engagement of error-correction reactions. Thus, in contrast to SAC's role as a safeguard mechanism for mitotic fidelity, removal of this checkpoint alleviates mitotic errors when sister chromatid cohesion is compromised (Silva, 2018).

In agreement with its 'safeguard' function, mitotic errors are often exacerbated by impairment of the SAC. These include defects associated with multiple centrosomes, defective microtubule assembly, or kinetochore structure. This study demonstrates that the opposite happens with cohesion defects. Absence of the SAC alleviated mitotic errors and improved mitotic fidelity after cohesion loss. Cells with a functional SAC undergo extensive chromosome shuffling and consequent genome randomization, whereas virtually no shuffling is observed in absence of the SAC. The detrimental nature of the SAC in the absence of cohesin is likely related to irreversibility of cohesion loss. Most mitotic defects can be corrected over time, but premature cohesin loss is an irreversible error and prolonging mitosis further enhances genome randomization (Silva, 2018).

Improved mitotic fidelity after cohesion loss in absence of the SAC is likely a consequence of slow kinetics of error-correction engagement coupled with a bias toward correct chromosome alignment. Several mechanisms are known to bias chromosome segregation toward the right orientation, including chromosome positioning, centromere geometry, bias on microtubule growth toward the kinetochores, and/or kinetochore-mediated microtubule nucleation. Among these, chromosome geometry is believed to facilitate bipolar attachment by facing one kinetochore to the opposite pole upon attachment. If so, what ensures geometric arrangement during the initial mitotic stages, even in the absence of cohesin? A possible mechanism enabling transient orientation of chromatids toward opposing poles is the incomplete resolution of sister chromatid intertwines. Yet residual catenation in metaphase chromosomes is unable to confer functional cohesion, as removal of cohesin is sufficient to induce immediate chromatid separation. Additional mechanisms may thus impair resolution of sister chromatids during early mitosis, in contrast to metaphase chromosomes. Spindle forces enhance decatenation, and thus resolution of DNA intertwines may only be achieved upon chromosome capture. Recent findings propose that efficient decatenation requires constant 'guiding action' from condensin I, whose maximal levels are observed on chromosomes in late metaphase or anaphase. This gradual condensin loading could limit full decatenation to later stages of mitosis, making residual catenation sufficient to allow transient pseudo-metaphase alignment that biases initial chromosome attachment toward biorientation. Although error prone, this process would be more accurate than total genome randomization resulting from chromosome shuffling (Silva, 2018).

Why cohesion loss is insufficient at triggering error correction during early mitosis remains to be addressed. This could be related to a partial tension state facilitated by pseudo-metaphase chromosomal configuration. Additionally, an intrinsic delayed action of error-correction machinery may further account for observed late shuffling onset. Indeed, slow kinetics or a lag time of Aurora-B-mediated chromosome detachment has been hypothesized in several theoretical studies with little experimental validation. Such intrinsic delay would solve the 'problem of initiation of biorientation' whereby initial low-tension interactions could survive such a tension-sensitive error correction (Silva, 2018).

Interestingly, the interplay between mitotic timing and sister chromatid cohesion has been previously reported in mammalian cells whereby extension of mitosis predisposes to sister chromatid cohesion defects. Cells arrested in mitosis for long periods were shown to display sister chromatid separation (referred as 'cohesion fatigue'). Moreover, defective sister chromatid cohesion was described to be synthetically lethal with impaired APC/C function in Warsaw breakage syndrome (WABS) patient-derived cells as well as several cancer cell lines with cohesion defects. The current observations demonstrate how reduction of mitotic timing is sufficient to rescue segregation defects associated with premature cohesin loss. Importantly, these experiments highlight the detrimental effect of the SAC upon cohesion defects. When sister chromatid cohesion is compromised and mitotic fidelity irreversibly affected, the SAC exacerbates mitotic errors in contrast to its canonical protective function (Silva, 2018).

DNA damage induces a kinetochore-based ATM/ATR-independent SAC arrest unique to the first meiotic division in mouse oocytes

Mouse oocytes carrying DNA damage arrest in meiosis I, thereby preventing creation of embryos with deleterious mutations. The arrest is dependent on activation of the spindle assembly checkpoint, which results in anaphase-promoting complex (APC) inhibition. However, little is understood about how this checkpoint is engaged following DNA damage. This study found that within minutes of DNA damage checkpoint proteins are assembled at the kinetochore, not at damage sites along chromosome arms, such that the APC is fully inhibited within 30 min. Despite this robust response, there is no measurable loss in k-fibres, or tension across the bivalent. Through pharmacological inhibition this study observed that the response is dependent on Mps1 kinase (see Drosophila Mps1), aurora kinase (see Drosophila Aurora B) and Haspin (see Drosophila Haspin). Using oocyte-specific knockouts this study found the response does not require the DNA damage response kinases ATM or ATR. Furthermore, checkpoint activation does not occur in response to DNA damage in fully mature eggs during meiosis II, despite the divisions being separated by just a few hours. Therefore, mouse oocytes have a unique ability to sense DNA damage rapidly by activating the checkpoint at their kinetochores (Lane, 2017).

Search PubMed for articles about Drosophila Mps1

Abrieu, A., et al. (2001). Mps1 is a kinetochore-associated kinase essential for the vertebrate mitotic checkpoint. Cell 106: 83-93. PubMed ID: 11461704

Fischer, M. G., Heeger, S., Häcker, U. and Lehner, C. F. (2004). The mitotic arrest in response to hypoxia and of polar bodies during early embryogenesis requires Drosophila Mps1. Curr. Biol. 14(22): 2019-24. PubMed ID: 15556864

Fisk, H. A., Mattison, C. P. and Winey, M. (2003). Human Mps1 protein kinase is required for centrosome duplication and normal mitotic progression. Proc. Natl. Acad. Sci. 100(25): 14875-80. PubMed ID: 14657364

Gilliland, W. D., Wayson, S. M. and Hawley, R. S. (2005). The meiotic defects of mutants in the Drosophila mps1 gene reveal a critical role of Mps1 in the segregation of achiasmate homologs. Curr. Biol. 15(7): 672-7. PubMed ID: 15823541

Gilliland, W. D., Hughes, S. E., Cotitta, J. L., Takeo, S., Xiang, Y. and Hawley, R. S. (2007). The multiple roles of mps1 in Drosophila female meiosis. PLoS Genet. 3(7): e113. PubMed ID: 17630834

Howell, B. J., et al. (2004). Spindle checkpoint protein dynamics at kinetochores in living cells. Curr. Biol. 14: 953-964. PubMed ID: 15182668

Jang, J. K., Rahman, T. and McKim, K. S. (2005). The kinesinlike protein Subito contributes to central spindle assembly and organization of the meiotic spindle in Drosophila oocytes. Mol. Biol. Cell 16: 4684-4694. PubMed ID: 16055508

Lane, S. I. R., Morgan, S. L., Wu, T., Collins, J. K., Merriman, J. A., ElInati, E., Turner, J. M. and Jones, K. T. (2017). DNA damage induces a kinetochore-based ATM/ATR-independent SAC arrest unique to the first meiotic division in mouse oocytes. Development 144(19): 3475-3486. PubMed ID: 28851706

Lince-Faria, M., et al. (2009). Spatiotemporal control of mitosis by the conserved spindle matrix protein Megator. J. Cell Biol. 184(5): 647-57. PubMed ID: 19273613

Liu, S. T. et al. (2003). Human MPS1 kinase is required for mitotic arrest induced by the loss of CENP-E from kinetochores. Mol. Biol. Cell 14: 1638-1651. PubMed ID: 12686615

O’Tousa, J. (1982). Meiotic Chromosome Behavior Influenced by Mutation - Altered Disjunction in Drosophila melanogaster Females. Genetics 102: 503-524. PubMed ID: 6816677

Palframan, W. J., et al. (2006). Anaphase inactivation of the spindle checkpoint. Science 313: 680-684. PubMed ID: 16825537

Pandey, R., Heeger, S. and Lehner, C. F. (2007). Rapid effects of acute anoxia on spindle kinetochore interactions activate the mitotic spindle checkpoint. J. Cell Sci. 120(Pt 16): 2807-18. PubMed ID: 17652159

Poss, K. D., et al. (2002). Mps1 defines a proximal blastemal proliferative compartment essential for zebrafish fin regeneration. Development 129: 5141-5149. PubMed ID: 12399306

Poss, K. D. et al. (2004). Germ cell aneuploidy in zebrafish with mutations in the mitotic checkpoint gene mps1. Genes Dev. 18: 1527-1532. PubMed ID: 15231734

Shonn, M. A., McCarroll, R. and Murray, A. W. (2000). Requirement of the spindle checkpoint for proper chromosome segregation in budding yeast meiosis. Science 289: 300-303. PubMed ID: 10894778

Shonn, M .A., Murray, A. L. and Murray, A. W. (2003). Spindle checkpoint component Mad2 contributes to biorientation of homologous chromosomes. Curr. Biol. 13: 1979-1984. PubMed ID: 14614824

Straight, P. D., Giddings, T. H. and Winey, M. (2000). Mps1p regulates meiotic spindle pole body duplication in addition to having novel roles during sporulation, Mol. Biol. Cell 11: 3525-3537. PubMed ID: 11029053

Stucke, V. M. Baumann, C. and Nigg, E. A. (2004). Kinetochore localization and microtubule interaction of the human spindle checkpoint kinase Mps1. Chromosoma 113: 1-15. PubMed ID: 15235793

Winey, M. and Huneycutt, B. J. (2002). Centrosomes and checkpoints: The MPS1 family of kinases. Oncogene 21: 6161-6169. PubMed ID: 12214245

date revised: 25 September 2023

Home page: The Interactive Fly © 2009 Thomas Brody, Ph.D.

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