Proteolytic degradation of mitotic regulatory proteins first requires these targets to be ubiquitinated. This is regulated at the level of conjugation of ubiquitin to substrates by the anaphase-promoting complex/cyclosome (APC/C) ubiquitin-protein ligase. Substrate specificity and temporal activity of the APC/C has been thought to lie primarily with its two activators, Cdc20/Fizzy and Cdh1/Fizzy-related. Reduction in the E2 ubiquitin-conjugating enzyme (UBC) of the E2-C family that is encoded by the Drosophila gene vihar (vih), by either mutation or RNAi, leads to an accumulation of cells in a metaphase-like state. Cyclin B accumulates to high levels in all mitotic vih cells, particularly at the spindle poles. Vihar E2-C is present in the cytoplasm of mitotic cells but also associates with centrosomes, and its own degradation is initiated at the metaphase-anaphase transition. Expression of destruction D box mutants of vihar in the syncytial embryo results in mitotic arrest at late anaphase. In contrast to hypomorphic mutants, Cyclin B is degraded at the spindle poles and accumulates in the equatorial region of the spindle. It is concluded that in Drosophila, the Vihar E2 UBC contributes to the spatiotemporal control of Cyclin B degradation that first occurs at the spindle poles. APC/C-mediated proteolysis of Vihar E2-C autoinactivates the APC/C at the centrosome before a second wave of proteolysis to degrade Cyclin B on the rest of the spindle and elsewhere in the cell (Mathe, 2004).

The ordered progression of cells through the division cycle is brought about by periodic series of protein modification events that involve cycles of phosphorylation that are mediated by multiple protein kinases and cycles of ubiquitination that lead to the periodic degradation of specific regulatory proteins. The ubiquitination of target proteins is achieved by three enzymes. The first ATP-dependent step is the activation of ubiquitin by the formation of a thio-ester bond between its C terminus and a cysteine residue in the activating enzyme E1 itself. In the second step, the ubiquitin is transferred as a thio-ester to a cysteine residue in a ubiquitin-conjugating enzyme (UBC) E2. In cooperation with an ubiquitin protein ligase, E3, the ubiquitin residue is then transferred to a lysine residue in the target protein. Polyubiquitinated proteins are targeted to the proteasome for their destruction. Proteolytic degradation is controlled in the cell cycle by two major classes of E3 enzyme: the Skp1 protein, Cullin, and F box (SCF) complex, which is required for the G1-S transition, and the anaphase-promoting complex/cyclosome (APC/C), which is functional during mitosis and G1. The APC/C catalyzes the ubiquitination of securin, an inhibitor of the protease that cleaves chromosome cohesion proteins. It is also responsible for degradation of the mitotic cyclins and of other regulatory molecules. The complex is comprised of 12 to 13 protein subunits and targets proteins that contain either of two types of destruction motif, the D box or the KEN box. APC/C activity is regulated in part by two related classes of WD40 repeat-containing proteins named after the Drosophila and budding yeast orthologs: Fizzy (Fzy) or Cdc20, which is required at the metaphase-anaphase transition, and Fizzy-related (Fzr) or Cdh1, which is effective toward the end of mitosisand into G1 to maintain mitotic cyclins at a low level. Activation of many APC/C functions may be delayed by the spindle assembly checkpoint that monitors microtubule attachment of and tension at kinetochores and signals Mad2 to complex with and inhibit the APC/C (Mathe, 2004 and references therein).

The mitotic cyclins are a major target of the APC/C, and in normal mitotic progression A-type cyclins are degraded ahead of the B-type cyclins. Experiments with stable forms of B-type cyclins in several organisms have shown their degradation to be required not at the time of chromatid separation but at later stages of mitosis. This is supported by real-time studies in Drosophila embryos that show stable Cyclin B1 (hereafter referred to simply as Cyclin B) functions to block spindle elongation at anaphase B, resulting in the oscillation of disjoined chromatids. Stable Cyclin B3 gave a late arrest in which anaphase and cytokinesis were completed, but chromosomes failed to decondense (Parry, 2001). Parry (2003) subsequently showed that the oscillation of disjoined chromatids resulted from the establishment of merotelic attachments of their kinetochores to both poles and that this was likely to be a consequence of the failure to release the Aurora B kinase from the kinetochore. It also had the consequence of blocking cytokinesis (Mathe, 2004 and references therein).

Early studies of Cyclin B behavior in syncytial Drosophila embryos had shown degradation to be incomplete, but these were complicated by the use of fixed preparations of embryos where the pattern of immunostaining depended upon fixation conditions. The use of GFP-tagged Cyclin B, however, has permitted time-lapse studies in both Drosophila and mammalian cells that have suggested that Cyclin B degradation begins first on the mitotic apparatus and then occurs subsequently in the cytoplasm (Huang, 1999; Clute, 1999). In Drosophila, Cyclin B is degraded on the spindle in a wave that spreads from the poles and then subsequently in the cytoplasm. Consistently, in mutant embryos derived from centrosome fall off (cfo) mothers, Cyclin B is degraded on the detached centrosomes but not on the acentrosomal spindles, as though the physical detachment presents a barrier to the wave of cyclin destruction (Wakefield, 2000). Degradation of spindle-associated Cyclin B has been attributed to APC/C associated with Fzy/Cdc20, and degradation of the cytoplasmic Cyclin B to Fzr/Cdh1, a protein that only appears to be active after cellularization (Raff, 2001). This led to the hypothesis that Fzy/Cdc20, localized on the kinetochores and centrosome prior to the metaphase-anaphase transition, might mediate the degradation of cyclin throughout the spindle once the metaphase checkpoint has been relieved at the kinetochore (Mathe, 2004 and references therein).

However, the above hypothesis does not satisfactorily account for how Fzy/Cdc20 might direct Cyclin B degradation to begin at the spindle poles rather than elsewhere on the spindle. The possibility that components of the ubiquitination pathway other than the APC/C and its associated proteins may contribute to determining the specificity of proteolytic degradation of proteins in mitosis by the APC/C was first raised by the finding of E2 enzymes that were specific for the mitotic cyclins (Aristarkhov, 1996; Yu, 1996). However, when the catalytic cysteine of the clam enzyme E2-C was changed to serine, this resulted in a dominant-negative form of the enzyme that was able to arrest mammalian cells in metaphase and inhibit destruction of both Cyclin A and Cyclin B (Townsley, 1997). Elimination of E2-C function through either mutation or RNA interference in Drosophila cells results in the accumulation of Cyclin B principally at the centrosomes and a characteristic delay of cells in a metaphase-anaphase-like state. Metazoan E2-C enzymes themselves contain putative destruction D boxes (Yamanaka, 2000). It is now shown directly that Drosophila E2-C is concentrated at the centrosome and that it is itself subject to cyclical degradation. Expression of a D box mutant form of the Vihar E2-C enzyme leads to mitotic defects in which Cyclin B is degraded at the spindle poles but not in the equatorial region of the spindle (Mathe, 2004).

This gene has been named vihar (Hungarian for storm) after the characteristic mutant phenotype in which chromosomes are scattered throughout the mitotic spindle. Similar reductions are observed in the levels of Vihar E2-C protein both in syncytial embryos derived from vih mutant mothers and S2 cells subjected to vih RNAi; both these treatments lead to comparable mitotic abnormalities. The majority of mitoses are delayed as metaphase figures in which chromosomes have undergone congression to the equator of the spindle. As seen with other mitotic mutants in Drosophila, many of these mitotic spindles lack centrosomes from one or both poles. Thus, this aspect of the phenotype cannot be attributed directly to reduction of the Vihar protein. The scattered chromosomes have the BubR1 checkpoint protein at their kinetochores, and the Aurora B passenger protein kinase appears not to have been transferred to the spindle, which does not adopted its characteristic late anaphase morphology. Such features have also been reported in cells arrested in mitosis due to the presence of nondegradable Cyclin B. Indeed, cells with reduced Vihar levels show pronounced increases in Cyclin B levels (Mathe, 2004).

The accumulation of Cyclin B at the centrosomes of vihar arrested cells is particularly striking and is an exaggeration of the association of the cdk1-Cyclin B complex with spindle poles that has been described in a wide number of organisms. The extent of degradation of Cyclin B varies during the embryonic development of Drosophila. Although the mitotic cyclins undergo extensive degradation at the metaphase-anaphase transition in cellularized Drosophila embryos and in tissues at later stages, they appear to persist throughout mitosis in the syncytial Drosophila embryo. There are successively increasing levels of Cyclin B degradation throughout the early syncytial cycles. This has been suggested to occur in restricted areas around the spindles as a result of observations of a gradient of the dephosphorylation of phospho-histone H3 along anaphase chromosomes, which is maximal near the spindle poles. Such a gradient has been interpreted as reflecting reduction of cdk1 activity near the spindle poles or centromere. Although this finding would probably now be interpreted as the activation of the protein phosphatase that opposes the B-type Aurora kinase that phosphorylates Histone H3, it nevertheless reflects a gradient of the activities of several enzymes associated with mitotic exit, which is initiated at the spindle poles. Support for this idea was provided by real-time imaging of GFP-tagged Cyclin B (Huang, 1999) that shows that its degradation begins at the spindle poles in cellularized embryos (Mathe, 2004).

Some indication of how Cyclin B degradation might be propagated along the spindle comes from studies of the localization of the Vihar E2-C protein and its own pattern of proteolysis. Immunolocalization experiments show not only that a considerable proportion of Vihar E2-C is associated with the centrosome, but also that it too undergoes degradation following the metaphase-anaphase transition. This was confirmed by the rapid destruction of accumulated Vihar E2-C protein following release of cells from a nocodazole block and also by the stabilization of the protein following treatment of embryos with a proteasome inhibitor. The Vihar centrosomal associated enzyme either dissociates or is directly degraded during anaphase, leaving a diffuse distribution of protein in the central part of the cell that is largely degraded upon mitotic exit. Thus, the spatiotemporal pattern of Vihar distribution is reminiscent of that described by Huang (1999) for Cyclin B. It suggests that degradation of the two proteins might be mediated by activated APC/C in similar subcellular compartments. However, Vihar E2-C appears to be more tightly localized to the centrosome in the syncytial embryo than Cyclin B, which has a more punctate distribution over astral microtubules and is more strongly associated with other regions of the spindle. Vihar also has a tighter distribution on the centrosome than the APC/C component Apc11 that clusters in a much larger 'cloud' of staining around the spindle poles as if associated with astral microtubules. Thus, the focus for APC/C activity may initially be the centrosome itself, where Vihar E2-C would not be rate limiting (Mathe, 2004).

Both of the Vihar D boxes appear to be required for efficient destruction of the Vihar protein, because mutations in either box alone result in defects in embryonic development, whereas expression of Vihar E2-C having mutations in both boxes gives a dominant mitotic phenotype. Previous studies with the mammalian counterpart of Vihar had shown that mutation of these sequences stabilized the protein in an in vitro assay, but it had not been possible to study effects in vivo (Yamanaka, 2000). Expression of the double D box mutant of Vihar leads to the preferential degradation of Cyclin B at the spindle poles and its continued presence in the equatorial region of the spindle, in contrast to the phenotype seen following downregulation of the enzyme when Cyclin B accumulates at the poles. How can this dominant effect be explained, and what is its relevance to wild-type Vihar function? Because the double D box mutant shows some reduction in catalytic activity (albeit measured in vitro against a nonphysiological substrate, a fragment of human Cyclin B), it is possible that the effects of downregulating Vihar function are being observed. However, it is noted that when a truly catalytically dead version of Vihar (mutated at its catalytic cysteine residue) is expressed in mitotic cells, the phenotype is similar to loss of Vihar function, metaphase arrest with accumulation of Cyclin B at the centrosomes. Nevertheless, the dominant effect of the Vihar double D box mutant is seen only at reduced concentration of the wild-type protein. In the whole organism, this is seen following reduction of the gene dosage of the wild-type allele, and in mitosis only once the wild-type protein has been partially destroyed. Thus, it may be that at this critical ratio of wild-type to D box mutant Vihar protein the dominant-negative effect comes into action to prevent the late pattern of Cyclin B degradation. However, it is felt that this explanation does not fully take into account the spatial allocation of Vihar to different subcellular compartments. There is no reason to suspect that at the onset of mitosis the ratio of wild-type to D box forms should vary whether associated with the centrosome, spindle, or cytoplasm. Thus, even though the ratio of wild-type to D box forms would first reduce at the poles, it should remain unchanged in other parts of the cell, where there should still be sufficient wild-type Vihar to give the potential for APC/C activity. Thus, the dominant-negative effect could be viewed from two perspectives, being a result either of reduced E2 activity or of continued E2 activity albeit at a reduced level, but in either event the mutant protein would persist at the spindle poles. The outcome of either of these viewpoints of the dominant effect might be similar: namely, the observed effect of preventing subsequent waves of Cyclin B proteolysis on the central part of the spindle and in the cytoplasm. Thus, the preferred hypothesis is that the continued presence of this nondegradable form of Vihar at the spindle poles could block mitotic progression, irrespective of its own level of E2 activity, by sequestering other rate-limiting components of the ubiquitination machinery. Conversely, in the wild-type situation, when the APC/C becomes competent at the metaphase-anaphase transition, the concentration of Vihar at the poles would direct ubiquitination of APC/C targets at this site, and its own polar destruction would be required to permit subsequent waves of proteolysis elsewhere in the cell (Mathe, 2004).

Perhaps the key to such a spatially regulated system lies in processes of microtubule-mediated transport. Raff (2001) has proposed that the degradation of spindle-associated Cyclin B might be mediated by APC/C associated with Fzy/Cdc20 and that Fizzy-related/Cdh1 directs more the degradation of Cyclin B in the cytoplasm. It is further suggested that once the metaphase-anaphase checkpoint is satisfied, flux of Fzy/Cdc20 from the kinetochore to the poles first mediates cyclin degradation. The finding that the Vihar E2-C is concentrated at the poles provides an explanation of why the Cyclin B degradation is initiated at the poles. The spatiotemporal pattern by which Vihar E2-C is then degraded suggests that it is ubiquitinated for degradation alongside its Cyclin B target once the APC/C has been activated in the vicinity of the spindle poles. Such autoregulatory inactivation of APC/C activity at the spindle poles might also release the APC/C or other regulatory components to mediate Cyclin B degradation within the central part of the spindle. This may be linked to a general influx of late mitotic regulators to the central spindle region. Indeed, plus end-directed motor proteins such as Pavarotti-KLP and several other regulators of cytokinesis begin to accumulate in the central spindle structure that forms at this time in preparation for cytokinesis. Once the ubiquitination process has been shut down at the poles, then similar autoinhibition could later operate upon Vihar E2-C in other regions of the cell following the subsequent wave of APC/C activity (Mathe, 2004).

Finally, it is noted that in vitro studies using Xenopus and clam extracts indicate that APC-mediated ubiquitination reactions are supported equally well by the Ubc4 and UBCx/E2-C enzymes. However, it has remained unclear whether these enzymes are redundant in vivo. A recent study from Seino (2003) in fission yeast, however, suggests that these two classes of E2 ubiquitin-conjugating enzymes are not functionally equivalent and have distinct roles in degrading the mitotic cyclin Cdc13. The phenotypes of cells deficient for Vihar E2-C enzyme indicate that it cannot be redundant with the Ubc4/5 family of E2 enzymes that remain functional in these cells. Thus, it remains of considerable interest to understand how different E2 family members cooperate in the APC/C-mediated destruction of mitotic cyclins and other mitotic regulatory proteins (Mathe, 2004).

It is concluded that the Vihar E2 UBC is concentrated at the spindle poles to facilitate the degradation of Cyclin B that first occurs at this site during the metaphase-anaphase transition. Vihar E2 UBC is also subject to ubiquitination and degradation by the APC/C. This results in the autoinactivation of APC/C-mediated ubiquitination at the spindle poles. Cyclin B degradation then takes place in the central spindle and other parts of the cell (Mathe, 2004).

GENE STRUCTURE

cDNA clone length - 1046

Bases in 5' UTR - 181

Exons - 2

Bases in 3' UTR - 328

PROTEIN STRUCTURE

Amino Acids - 178

Structural Domains

The amino acid sequence of Vihar reveals that it belongs to the E2 family of ubiquitin-conjugating enzymes. Within this family, Vihar shows more than 50% amino acid identity to E2s that are involved in the degradation of mitotic cyclins, the E2-Cs. The most highly conserved region of the Vihar protein is the catalytic UBC domain (amino acids 105-119), within which lies a specific cysteine residue that is required for the ubiquitin-E2 thio-ester formation. The protein also contains two putative destruction boxes (D1 and D2) at almost identical positions as those in the mammalian E2-Cs. Vihar was conformed to be able to function as an E2 enzyme; the bacterially expressed protein is able to mediate the APC/C-dependent ubiquitination of Cyclin B. Mutant forms of Vihar were constructed for expression in E. coli; in one the catalytic cysteine residue was mutated to alanine, and in a second basic arginine and leucine amino acids in its two putative D boxes (RxxL) were mutated to alanine residues. Whereas the former showed no catalytic activity in vitro, the latter was still able to generate multiubiquitinated forms of Cyclin B, although at a reduced efficiency compared to the wild-type protein (Mathe, 2004).

date revised: 5 March 2005

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

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