BIOLOGICAL OVERVIEW

Drosophila Pimples (Pim) and Three rows (Thr) are novel proteins unrelated to other known proteins of yeast or mammals. Both are required for sister chromatid separation in mitosis. Chromosomes are held together by 'Cohesin', a highly conserved multiprotein complex thought to be the primary effector of sister-chromatid cohesion in all eukaryotes, including Drosophila. Cohesin complexes in budding yeast hold sister chromatids together from S phase until anaphase, but in metazoans, cohesin proteins dissociate from chromosomes and redistribute into the whole cell volume during prophase, well before sister chromatids separate. Proteolytic cleavage of one of cohesin's subunits by a protease separin may trigger sister separation at the onset of anaphase (reviewed by Dej, 2000 and Nasmyth, 2000).

Pimples accumulates during interphase and is degraded rapidly during mitosis. This degradation is dependent on a destruction box similar to that of B-type cyclins. Nondegradable Pim with a mutant destruction box can rescue sister chromatid separation in pim mutants but only when expressed at low levels. Higher levels of nondegradable Pim, as well as overexpression of wild-type Pim, inhibit sister chromatid separation. Moreover, cells arrested in mitosis before sister chromatid separation (by colcemid or by mutations in fizzy/CDC20) fail to degrade Pim. Thus, although not related by primary sequence, Pim has intriguing functional similarities to the securin proteins of budding yeast, fission yeast, and vertebrates. Whereas these securins are known to form a complex with separins (see Drosophila Separase), PIM associates in vivo with Thr, which does not contain the conserved separin domain. The following overview suggests that Pim may represent a securin protein involved in the regulation of a Drosophila separin (Leismann, 2000). Analysis of the functions of Pim and Thr raise important questions about the evolutionary conservation of proteins involved in sister chromtid exchange, focusing on the extent their divergence coincident with the conservation of function.

Pairs of sister chromatids are generated during the S phase of the eukaryotic cell division cycle. Sister chromatids remain paired throughout the G2 phase and during the initial phase of mitosis (prophase) while chromatin is condensed and the spindle is assembled. However, the cohesion between sister chromatids is ultimately destroyed at the metaphase-anaphase transition allowing their segregation to opposite poles. Considerable progress has been made in understanding the molecular basis of cohesion and separation of sister chromatids (for review, see Zachariae, 1999 and Nasmyth, 2000). Cohesion is known to be dependent on the binding of the cohesin protein complex to nascent sister chromatids during S phase. Separation of sister chromatids in budding yeast mitosis requires the proteolytic processing of the cohesin subunit Scc1p/Mcd1p during the metaphase-anaphase transition (Uhlmann, 1999). In vertebrates, cohesin complexes dissociate from chromosomes already during prophase concomitant with chromatin condensation and well before the onset of sister chromatid separation (Losada, 1998; Darwiche, 1999). Moreover, cleavage of the Drosophila Scc1p homolog (Rad21) during prophase is not detectable in Drosophila (S. Heidmann, unpubl., reported in Leismann, 2000). The possibility is not excluded that residual cohesin complexes might persist in particular in the centromeric region of vertebrate chromosomes. The final separation of sister chromatids in higher eukaryotes, therefore, might also result from Scc1p cleavage during the metaphase-anaphase transition (Leismann, 2000 and references therein).

This hypothesis of a conserved mechanism of sister chromatid separation in eukaryotes is supported by findings concerning the role of the separin and securin proteins (reviewed Nasmyth, 2000). The separins (Esp1p, Cut1, BimB) were implicated originally in mitosis based on genetic analyses in fungi. Homologous genes have been detected recently in plant and animal species. All these separins share a conserved carboxy-terminal domain, the separin domain. The budding yeast separin Esp1p is known to be required for Scc1p cleavage and sister chromatid separation (Uhlmann, 1999). Separins are thought to be activated only during the metaphase-anaphase transition. Premature activation of separins is prevented by securin proteins that accumulate during interphase and bind to the separins. The budding yeast securin Pds1p forms a complex with Esp1p. The fission yeast securin Cut2 binds to Cut1. In vertebrates, the protein encoded by the pituitary tumor transforming gene (PTTG) associates with a protein containing the conserved separin domain (Zou, 1999). All these securins (Pds1p, Cut2, PTTG) share essentially no sequence similarity except for the presence of at least one destruction box, a nine amino acid consensus motif [RX(A or V or L)LGXXXN] originally defined in B-type cyclins. Securins are therefore degraded rapidly during the metaphase-anaphase transition similar to the mitotic cyclins. Securin proteins with mutations in the destruction box fail to be degraded and inhibit sister chromatid separation in yeast and in Xenopus extracts (Cohen-Fix, 1996; Funabiki, 1996, 1997 and Zou, 1999, all cited in Leismann, 2000).

Mitotic proteolysis of destruction box proteins occurs after polyubiquitination resulting from the activation of a special ubiquitin ligase known as anaphase-promoting complex/cyclosome (APC/C). APC/C activation, therefore, is a crucial step in the regulation of the metaphase-anaphase transition (for review, see Zachariae and Nasmyth, 1999). This activation process is not yet fully understood. However, it is clear that the WD-40 repeat proteins Fizzy/Cdc20p and Fizzy-related/Hct1p/Cdh1p (Drosophila Fizzy-related is Retina aberrant in pattern) play important roles in APC/C regulation. These proteins bind to the APC/C in different cell cycle phases and respond to different regulatory inputs. While Drosophila Fizzy-related is known to be essential for the degradation of mitotic cyclins in G1, Fizzy is required for cyclin degradation and sister chromatid separation during mitosis (Sigrist, 1995; Sigrist, 1997). The dependency of sister chromatid separation on Cdc20p function has been explained in budding yeast by the finding that Cdc20p is required for the degradation of the securin Pds1p (Visintin, 1997; Lim, 1998 and Shirayama, 1999). Fizzy/Cdc20p is inactivated in the presence of unattached kinetochores and spindle damage by a mitotic checkpoint pathway which results in the binding of the inhibitor Mad2p to the Fizzy/Cdc20p-APC/C complex. This checkpoint pathway therefore assures that sister chromatid separation and exit from mitosis occur only when all chromosomes have acquired the correct bipolar orientation within a functional spindle (Leismann, 2000 and references).

With the exception of securins, all the components involved in the control of sister chromatid separation that have been introduced above are highly conserved in eukaryotes. Interestingly, two nonconserved Drosophila genes, pimples (pim) and three rows (thr), are both required specifically for sister chromatid separation during mitosis (Smith, 1993; D'Andrea, 1993; Stratmann and Lehner, 1996). Pimples protein shares extensive functional similarities with securin proteins and in particular with Cut2 from fission yeast. Moreover, Pim is found in a complex with Three rows protein (THR). These results indicate that the regulation of sister chromatid separation in Drosophila involves non-conserved securin-like proteins that associate with proteins lacking the evolutionary conserved separin domain (Leismann, 2000).

Mitotic degradation of Cyclin A, Cyclin B, and Cyclin B3 requires Fizzy/Cdc20p, an activator of APC/C-dependent ubiquitination. To evaluate whether Fizzy is also involved in Pim degradation during mitosis, the consequences of UAS-pim-myc expression in fizzy mutants was analyzed. The maternal fizzy contribution present in fizzy mutants is sufficient for progression through all of the 16 embryonic divisions in the dorsal epidermis when UAS-pim-myc is not expressed. However, when UAS-pim-myc is expressed, sister chromatid separation is found to be inhibited in the dorsal epidermis of fizzy mutants during mitosis 16, while exit from this mitosis 16 still occurs. Importantly, in contrast to the results observed in wild-type embryos, expression of just one UAS-pim-myc copy is already sufficient for inhibition of sister chromatid separation in the dorsal region of fizzy mutants and results in a phenotype that is only observed in wild-type embryos when two UAS-pim-myc copies are expressed (Leismann, 2000).

In the ventral region of fizzy homozygotes, the maternal fizzy contribution is not sufficient to allow completion of mitosis 16. Therefore, a large fraction of ventral cells become arrested during metaphase 16 in fizzy mutants. When UAS-pim-myc is expressed in fizzy mutants, very strong anti-myc labeling is observed in the arrested cells of the ventral region. This labeling is much more intense than in the dorsal UAS-pim-myc expressing cells that are not arrested. The persistence of Pim-myc during metaphase arrest resulting from lack of fizzy function is also observed when expression is directed at lower levels by transgenes under the control of the pim⁺ regulatory region. Moreover, immunoblotting experiments have confirmed that the endogenous Pim protein is also stabilized in fizzy homozygotes. It is concluded, therefore, that fizzy is required for Pim degradation during mitosis (Leismann, 2000).

Spindle defects result in a mitotic checkpoint arrest during which sister chromatids do not separate, possibly because Pim is not degraded. To evaluate whether Pim is stable during a mitotic checkpoint arrest, UAS-pim-myc-expressing embryos were treated with colcemid and mock-treated embryos were used as control. These embryos were subsequently labeled with a DNA stain to identify arrested cells and with anti-myc antibodies to monitor the presence of Pim-myc. The embryos were also labeled with an antibody against Cyclin A, which is known to be degraded in colcemid-arrested cells in contrast to Cyclin B. Pim-myc remains clearly detectable in mitotic domains of arrested cells with condensed chromosomes and without Cyclin A labeling. These observations demonstrate that Pim is not degraded in cells arrested by the spindle checkpoint pathway (Leismann, 2000).

The Pim persistence in cells arrested by colcemid or lack of fizzy, as well as the inhibition of sister chromatid separation resulting from UAS-pim^dba-myc and UAS-pim-myc expression, are consistent with the notion that sister chromatid separation is strictly dependent on Pim degradation during mitosis. However, the experiments with UAS-pim^dba-myc and UAS-pim-myc involved overexpression. To analyze the effects of physiological levels of Pim^dba-myc, a transgene was constructed with the pim⁺ regulatory region directing Pim^dba-myc expression (gpim^dba-myc). Interestingly, transgenic lines could be established indicating that expression of a single gpim^dba-myc copy is tolerated in a pim⁺ background. However, when present in two copies, transgene insertions result in complete lethality (five out of eight lines) or severe morphological abnormalities (rough eyes, notched wings, sterility) in rare escapers (three out of eight lines). The analysis of heterozygous combinations of different transgene insertions indicates that these phenotypes are not caused by transgene insertion position effects. The phenotypes indicate clearly that Pim^dba-myc is highly toxic (Leismann, 2000).

The fact that transgenic lines with gpim^dba-myc insertions could be isolated suggests that sister chromatid separation is not absolutely dependent on complete Pim degradation during each mitosis. However, gpim^dba-myc expression might occur only at very low levels as a result of a selection against insertions generating normal expression levels during transgene establishment. Moreover, wild-type Pim might compete with Pim^dba-myc and thereby protect cells. Therefore, gpim^dba-myc expression levels were addressed in immunoblotting experiments and the consequences of gpim^dba-myc expression was analyzed in pim mutants (Leismann, 2000).

The insertion gpim^dba-myc II.5, which results in lethality when homozygous, was found to result in expression levels that were only ~25% lower as those of the endogenous locus. In these experiments, protein products resulting from early zygotic expression during <2 embryonic cell cycles were compared. Comparison of protein levels that had accumulated during this brief phase was chosen, since the difference in Pim-myc and Pim^dba-myc levels is likely to increase with every cell cycle due to the differential stability during mitosis. Moreover, the maternal pim⁺ contribution is known to be exhausted in pim mutants at this stage. Thus the consequences of gpim^dba-myc expression on progression through mitosis in the absence of wild-type pim⁺ were also analyzed at this stage. For this analysis, gpim^dba-myc II.5 was recombined with a mutant pim allele (pim¹) that abolishes expression from the endogenous locus. Analysis of mitosis 15 in pim¹/pim¹, gpim^dba-myc embryos and in pim¹, gpim^dba-myc/pim¹, gpim^dba-myc embryos indicates that sister chromatid separation occurred almost normally. Moreover, the same observations were also made during mitosis 16. As in wild-type embryos, anaphase and telophase figures were observed readily in these embryos in cells lacking Cyclin B labeling. However, a significant fraction of anaphase and telophase figures (~10%) had chromatin bridges, suggesting that sister chromatid separation is not always normal. Moreover, gpim^dba-myc fails to rescue the development of pim¹ homozygotes to the adult stage, whereas the lethality associated with pim¹ is prevented by gpim-myc. Nevertheless, the very significant rescue of sister chromatid separation during the embryonic divisions obtained with gpim^dba-myc in pim¹ mutants, demonstrates that Pim^dba-myc can still provide some positive function required for sister chromatid separation. In addition, coimmunoprecipitation experiments indicate that Pim^dba-myc is associated with Thr. The dba mutation therefore appears to interfere specifically with mitotic degradation. The observation that sister chromatid separation is not inhibited in the presence of near physiological levels of nondegradable Pim^dba-myc argues strongly that sister chromatid separation is likely to be controlled by other mechanisms operating in addition to Pim degradation (Leismann, 2000).

It is concluded that Pim is degraded during the metaphase-anaphase transition. This mitotic degradation of Pim is dependent on a destruction box motif that deviates at the first invariant position of the hitherto established destruction box consensus. While all previously characterized destruction boxes start with an arginine, a lysine is found in the Pim motif. However, this Pim variant can replace the destruction box of Cyclin B. Moreover, Fizzy, an activator of the APC/C that is required for the degradation of mitotic cyclins, is also required for Pim degradation. It is assumed, therefore, that Pim is degraded by the proteasome after APC/C-dependent polyubiquitination just like the mitotic cyclins (Leismann, 2000).

Pim degradation during mitosis appears to be an important step for sister chromatid separation. Overexpression of wild-type Pim and nondegradable Pim results in a complete inhibition of sister chromatid separation. Sister chromatid separation is equally defective in the absence of Pim (Stratmann, 1996). Pim therefore can act as both an activator and an inhibitor of sister chromatid separation (Leismann, 2000).

Inhibitors of sister chromatid separation that are degraded during the metaphase-anaphase transition by the APC/C pathway, and that function to a variable extent as activators of sister chromatid separation, have been described previously in yeast and vertebrates. These securin proteins (S. cerevisiae Pds1p, S. pombe Cut2, vertebrate PTTG) have an additional property in common. They all bind to proteins containing a conserved carboxy-terminal separin domain (S. cerevisiae Esp1p, S. pombe Cut1, vertebrate Esp1p). The separin proteins play a crucial role for sister chromatid separation. The functional characterization of budding yeast separin Esp1p has suggested that it may function as a protease that cleaves a cohesin subunit and thereby causes the dissolution of sister chromatid cohesion at the metaphase-anaphase transition (Uhlmann, 1999; Nasmyth, 2000). In addition, separin proteins might also be involved in the regulation of spindle function (Yanagida, 2000). Premature activation of separin activity is restricted by the securin proteins (Ciosk, 1998; Kumada, 1998 and Uhlmann, 1999, all cited in Leismann, 2000).

It is not known whether Pim binds to a protein with a separin domain. The Drosophila genome sequence predicts the existence of a protein with a separin domain (see Separase). Interestingly, this Drosophila separin homolog exhibits less sequence conservation than all the other known separin proteins from fungi, plants, nematodes, and vertebrates. The Drosophila protein is highly divergent in one of the conserved motifs within the carboxy-terminal separin domain and has a relatively small amino-terminal domain (H. Jäger, S. Heidmann and C.F. Lehner, unpubl., cited in Leismann, 2000). The amino-terminal regions of separin proteins generally show very little sequence conservation, but the fission yeast securin has been shown to bind to this nonconserved region. Although the separin proteins share at least a related carboxy-terminal domain, securin proteins do not display significant sequence similarity. The only conserved feature of the securins is the distribution of charged residues. The amino-terminal region is highly basic and the carboxy-terminal region highly acidic. This charge distribution is also present in Pim. It is therefore entirely possible that Pim represents a securin protein involved in the regulation of a Drosophila separin (Leismann, 2000).

Although it is not yet known whether Pim binds to a separin protein, Pim definately associates with Thr in vivo. Like Pim, Thr is also required for sister chromatid separation and shows no significant similarity to known proteins (D'Andrea, 1993). A further detailed characterization of the Pim-Thr complex is underway, including an analysis of its relationship to separin complexes. Three main hypotheses will have to be addressed (Leismann, 2000):

1.	The separin gene might have broken apart during the evolution of Drosophila melanogaster resulting in the thr gene encoding the nonconserved amino-terminal region and a distinct gene encoding the conserved carboxy-terminal separin domain, which might also be part of the Pim-Thr complex. The Pim-Thr complex might therefore be largely equivalent to the securin/separin complex.
2.	Instead of playing the role of the amino-terminal separin region, Thr might be a novel separin-associated protein. Unknown separin-associated proteins in addition to the known securins were revealed by affinity purification in Xenopus (Zou, 1999) but not in budding yeast (Ciosk, 1998).
3.	The Pim-Thr complex might be distinct from securin/separin complexes. It is emphasized that the role of higher eukaryote separin proteins have not yet been studied in detail and that the vertebrate homolog of the budding yeast cohesin subunit Scc1p dissociates from chromatin already during prophase, presumably as a requirement for chromatin condensation (Nasmyth, 2000). It remains a possibility therefore, that additional mechanisms have evolved in higher eukaryotes to maintain sister chromatid cohesion until the metaphase-anaphase transition, in particular in the centromeric region. The Pim-Thr complex might be involved in the dissolution of this residual cohesion maintained in the centromeric region of higher eukaryote mitotic chromosomes. Maintenance of cohesion in the centromeric region is also of particular importance during the first meiotic division and a gene specifically required for this maintenance has been identified in Drosophila. Interestingly, this gene (Mei-S332) has no obvious homologs in other species, and it has been proposed to be involved in maintaining cohesion not only during the first meiotic division but also during mitotic divisions (Tang, 1998).

Apart from their shared functions (inhibition of sister chromatid separation, separin binding, APC/C substrate) the securins Pds1p, Cut2, and PTTG differ with regard to their role as positive regulators of sister chromatid separation and their involvement in checkpoint mechanisms. The following comparisons indicate that Pim is functionally most similar to fission yeast Cut2. Both proteins provide a positive and essential function. They are absolutely required for sister chromatid separation (Uzawa, 1990 and Stratmann, 1996). In contrast, budding yeast Pds1p is clearly not required for sister chromatid separation in unperturbed cells. Pds1p is only essential at high temperatures and in the presence of DNA damage, lagging chromosomes and spindle damage. Moreover, while Cut2 appears to function only in mitotic checkpoint control, Pds1p has been implicated in both DNA damage and mitotic checkpoint pathways. Overexpression of nondegradable mutant Pds1p inhibits both sister chromatid separation and exit from mitosis in budding yeast. In contrast, only the former process appears to be blocked by nondegradable mutant Cut2 in fission yeast. This differential involvement in checkpoint regulation might reflect a fundamental physiological difference between budding and fission yeast. DNA damage causes fission yeast (and animal cells) to block entry into mitosis primarily by preventing Cdk1 activation. In contrast, DNA damage inhibits the metaphase-anaphase transition in budding yeast and Pds1p is of major importance for this DNA damage checkpoint arrest. It is not known whether Pim and vertebrate PTTG are involved in a DNA damage checkpoint pathway. However, based on the effects caused by the expression of nondegradable mutant proteins, Pim and vertebrate PTTG appear to be more similar to Cut2 than to Pds1p. Nondegradable PTTG and Pim inhibit only sister chromatid separation but not the degradation of mitotic cyclins and exit from mitosis (Zou, 1999 and Leismann, 2000 and references therein).

The mitotic checkpoint pathway delays the metaphase-anaphase transition in the presence of unattached kinetochores, and prevents the activation of APC/C by Fizzy/Cdc20p (Zachariae, 1999). As a consequence, mitotic cyclins and securins persist in the arrested cells. In budding yeast, the persistence of Pds1p has been demonstrated to be required for the inhibition of sister chromatid separation by elegant and conclusive genetic experiments involving cells lacking PDS1. Although not proven, the persistence of fission yeast Cut2 and vertebrate PTTG in checkpoint arrested cells is also thought to be responsible for the inhibition of sister chromatid separation. Pim persists in mitotic checkpoint-arrested cells as well. As in the case of Cut2, however, the essential positive role of Pim in sister chromatid separation makes it impossible to demonstrate that this Pim persistence is responsible for the inhibition of sister chromatid separation in the same elegant way realized in budding yeast. It is clear, however, that Pim levels are of crucial importance for sister chromatid separation. Modest overexpression of nondegradable Pim leads to a complete inhibition of sister chromatid separation. Moreover, higher but still relatively modest levels of wild-type Pim overexpression (~fivefold) inhibit as well. In this context, it is considered very likely that overexpression of the vertebrate securin PTTG results in chromosome instability, explaining its oncogenic potential (Leismann, 2000 and references therein).

Even though these findings demonstrate clearly the importance of Pim degradation, they cannot answer the question whether Pim protein persistence is responsible for the inhibition of sister chromatid separation in cells arrested by the spindle checkpoint. In this case, one would expect physiological levels of nondegradable Pim to inhibit sister chromatid separation completely. However, most cells with near physiological levels of mutant, nondegradable Pim protein (~75% of wild type), instead of normal Pim, appear to separate sister chromatids without major problems. Although not excluded, it appears unlikely, therefore, that the persistence of Pim protein during a mitotic checkpoint arrest is solely responsible for the block of sister chromatid separation. Pim persistence might be only one of several measures that cooperatively prevent premature sister chromatid separation in the presence of spindle damage. However, it remains to be shown that the myc epitopes present at the carboxyl terminus of the nondegradable Pim protein expressed in these experiments do not reduce the anaphase inhibitor function of Pim (Leismann, 2000).

The almost normal progression through mitosis in the presence of near physiological levels of nondegradable Pim suggests that the onset of sister chromatid separation is not determined exclusively by the kinetics of Pim degradation. No increase in the fraction of cells with metaphase plates has been detected in the presence of nondegradable Pim and careful examination by confocal microscopy does not reveal any residual partial degradation of the mutant Pim protein. The notion that the timing of sister chromatid separation during mitosis can be controlled by pathways that are independent of Pim degradation is also supported by the observation that expression of nondegradable Cyclin A clearly delays the metaphase-anaphase transition (Sigrist, 1995) even though it does not appear to result in a delay of Pim degradation (Leismann and Lehner, unpubl., cited in Leismann, 2000). The timing of sister chromatid separation under normal conditions in budding yeast, when spindle or DNA damage checkpoints are not activated, is also not controlled by Pds1p degradation, because it occurs with normal kinetics in cells lacking Pds1p (Leismann, 2000).

It is concluded that Pim levels that are controlled by mitotic degradation are of crucial importance for sister chromatid separation. Pim therefore shares extensive similarities with securin proteins. Although its role in the regulation of Drosophila separin remains to be analyzed, it is clear that it associates with Thr, a protein that is equally important for sister chromatid separation and that does not contain a separin domain (Leismann, 2000).

GENE STRUCTURE

cDNA clone length - 1129

Exons - 3

PROTEIN STRUCTURE

Amino Acids - 199

Structural Domains

Neither Pim nor Thr share significant sequence similarities with known proteins, and their biochemical function is not known (Stratmann, 1996).

date revised: 15 June 2001

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.