Protein Interactions

To investigate whether the putative Drosophila securin Pim can bind to Sse, the yeast two-hybrid system was used. A strong interaction between Pim and Sse was seen. Interestingly, a mutant Pim protein with a small internal deletion (amino acids 110-114) failed to bind to Sse. This deletion was identified in pim2, a Drosophila allele that results in an amorphic phenotype (Stratmann, 1996). Whereas Pim2 fails to interact with Sse, it binds to a Thr fragment (amino acids 1-933) just like wild-type Pim. The deletion of amino acids 110-114 therefore abolishes specifically the binding to Sse and does not result in destabilization or complete misfolding of the mutant Pim2 protein. It is concluded that the interaction of Pim and Sse is likely to be functionally significant (Jäger, 2001).

The behavior of Pim2 suggested that different domains of Pim mediate the binding to Sse and Thr. To evaluate this notion, Pim fragments were analyzed in additional two-hybrid experiments. Like other securin proteins, Pim is composed of a basic N-terminal and an acidic C-terminal domain. The N-terminal domain interacts with Sse, whereas an interaction with Thr 1-933 is barely detectable. Conversely, the C-terminal domain interacts with Thr 1-933, but not with Sse. These results strongly indicate that Sse and Thr bind to different Pim domains (Jäger, 2001).

Experiments with Sse fragments indicate that Pim binds to the N-terminal regions of Sse. The interaction of Pim with full-length Sse or with a Sse fragment comprising amino acids 1-467 appears to be stronger than with a shorter Sse fragment (amino acids 1-247). It is assumed that region 1-247 of Sse is sufficient for Pim binding, but the region 248-467 of Sse further strengthens this interaction. Clearly, the conserved C-terminal region of Sse is not required for the interaction with Pim (Jäger, 2001).

Experiments with Thr fragments indicate that Pim binds to the N-terminal region of Thr. Region 1-476 of Thr is sufficient for Pim binding. An interaction between Pim and the full-length Thr protein was not observed. The considerable size of full-length Thr 1-1379 might preclude expression of sufficiently high levels and/or entry into the nucleus (Jäger, 2001).

Finally, tests were performed for a direct interaction between Thr and Sse. The Thr fragment 1-933 interacts with full-length Sse, Sse 1-247, and Sse 1-467. These results therefore raise the possibility that Thr and Pim bind to the same region of Sse and thus might be competing in vivo. Moreover, the interactions between Thr and Sse appear to be weaker than the interactions between Pim and Sse, since in the former case, only one of the two reporter genes was activated. It was not possible to define the region in Thr that mediates Sse binding in more detail, because the shorter constructs Thr 1-476, Thr 208-933, and Thr 477-933 failed to bind to Sse (Jäger, 2001).

Taken together, the results of the two-hybrid experiments show that any two of these three proteins, Pim, Thr, and Sse can interact with one another, independent of the third (Jäger, 2001).

To confirm that Pim, Thr, and Sse interact in vivo, coimmunoprecipitation experiments were performed. Two antibodies against Sse were raised and affinity purified. In embryo extracts, these antibodies detect a prominent band at 75 kD, which comigrates with Sse translated in vitro. This band was also detected in immunoprecipitates that were isolated with an anti-myc antibody from extracts of embryos expressing Pim or Thr fused with myc epitope tags. Control experiments indicated that coimmunoprecipitation of Sse with Pim-myc and Thr-myc is specific. This specific association was also observed when the antibodies against Sse were used for immunoprecipitation followed by immunoblotting with anti-myc. Taken together, these results clearly show that Sse associates in vivo with Pim and Thr (Jäger, 2001).

The interaction between Thr and Sse observed in the yeast two-hybrid experiments was weak, raising the possibility that it might not be sufficiently strong to allow formation of Thr-Sse complexes in vivo. Pim might therefore be required to bring Thr and Sse together. To evaluate whether Thr-Sse complexes can be formed in the absence of Pim in vivo, gUAS-thr-myc (which allow expression of myc-tagged products under the control of the normal genomic regulatory regions) and UAS-HA-Sse were expressed in late embryos when endogenous Pim and Thr levels are very low. After precipitation of HA-Sse with an anti-HA antibody, Thr-myc as well as minor amounts of Pim were detected in the immunoprecipitates. Quantitative immunoblotting revealed that Thr-myc is present in an at least fivefold molar excess over Pim in these immunoprecipitates. Because Thr does not form oligomers (Leismann, 2000), it is concluded that Thr and Sse can associate in the absence of Pim (Jäger, 2001).

To investigate whether Pim is able to bind to Sse in the absence of Thr in vivo, a complementary experiment was performed by expressing UAS-HA-Sse and UAS-pim-myc in late embryos. Surprisingly, in contrast to the results of the yeast two-hybrid experiments, only minor amounts of Pim-myc could be coprecipitated with Sse. However, simultaneous coexpression of gUAS-thr-myc, UAS-HA-Sse, and UAS-pim-myc results in the recovery of significant Pim levels. It is estimated that <10% of Pim is precipitated in the absence of Thr when compared with the amount precipitated in the presence of Thr. Significantly, Pim2-myc does not form a stable complex with HA-Sse, even in the presence of Thr. These results indicate that Pim, Thr, and Sse form a trimeric complex in vivo, and that Sse is not sufficient to recruit Pim in the absence of Thr. This notion was supported by the behavior of a Thr deletion mutant (Thr 445-1379-myc), which was lacking the region required for the two-hybrid interaction with Pim. When this mutant protein was expressed from a transgene under the control of the normal thr regulatory region and immunoprecipitated from embryo extracts, almost no coimmunoprecipitation of Pim was detected. However, Sse was readily detected in the immunoprecipitates, confirming that Thr and Sse can form a complex without Pim. Quantitative immunoblotting experiments indicated that Sse associates with Thr 445-1379-myc with at least 50% efficiency when compared with its binding to full-length Thr, whereas binding of Pim to Thr 445-1379-myc was reduced to <5%. On the basis of the results of two-hybrid experiments, in which Sse interacts strongly with Pim, the Sse present in the Thr 445-1379-myc immunoprecipitates would be expected to result in coimmunoprecipitation of Pim as well. However, the absence of Pim in the immunoprecipitates suggests that Pim cannot join Sse and Thr in a trimeric complex when it is not bound by Thr. Control experiments with Thr-myc versions that contained the N-terminal Pim-binding region show that coimmunoprecipitation of Pim along with Sse can be readily detected in these cases (Jäger, 2001).

Thr 1-478-myc does not form complexes with Pim in vivo, whereas Thr 1-476 and Pim associate in yeast two-hybrid experiments. This discrepancy might result from different positions of the fused tags. Whereas in the two-hybrid experiments, the GAL4-binding domain was fused to the N terminus, the 10 myc epitope tags were fused to the C terminus of the Thr fragment analyzed in Drosophila embryos. The C-terminal myc tags, therefore, might interfere with Pim binding to Thr 1-478-myc (Jäger, 2001).

The C-terminal fragment Thr 932-1379-myc does not bind Pim or Sse, as expected from the yeast two-hybrid analysis. This fragment is present at very low levels in the embryo extracts as determined for four independent transgenic lines. This low abundance is presumably due to protein instability, since all genomic Thr constructs contained identical 5'- and 3'-noncoding regulatory sequences. No coimmunoprecipitation of Pim or Sse was detected when loading was adjusted to compensate for the low abundance of Thr 932-1379-myc (Jäger, 2001).

In summary, the behavior of Thr fragments expressed in embryos during the proliferative stages extended the findings resulting from yeast two-hybrid experiments. It is observed that the Thr-Sse interaction is not necessarily mediated by Pim. Furthermore, the results show that the binding of Pim to Sse requires Thr, strongly suggesting the existence of trimeric Pim-Thr-Sse complexes in vivo (Jäger, 2001).

Sister-chromatid separation in mitosis requires proteolytic cleavage of a cohesin subunit. Separase, the corresponding protease, is activated at the metaphase-to-anaphase transition. Activation involves proteolysis of an inhibitory subunit, securin, following ubiquitination mediated by the anaphase-promoting complex/cyclosome. In Drosophila, the securin Pimples (Pim) associates not only with Separase (Sse), but also with an additional protein, Three rows (Thr). Thr is cleaved after the metaphase-to-anaphase transition. Thr cleavage occurs only in functional Sse complexes and in a region that matches the separase cleavage-site consensus. Mutations in this region abolish mitotic Thr cleavage. These results indicate that Thr is cleaved by Sse. Expression of noncleavable Thr variants results in cold-sensitive maternal-effect lethality. This lethality can be suppressed by a reduction of catalytically active Sse levels, indicating that Thr cleavage inactivates Sse complexes. Thr cleavage is particularly important during the process of cellularization, which follows completion of the last syncytial mitosis of early embryogenesis, suggesting that Drosophila separase has other targets in addition to cohesin subunits (Herzig, 2002).

Immunolabeling has revealed that Thr is partially degraded after the metaphase-to-anaphase transition, similar to Pim. However, the mitotic degradation of Pim and Thr is mechanistically and functionally distinct. Mitotic degradation of Pim is dependent on the presence of a destruction box (D-Box) and on Fizzy-APC/C, which promotes ubiquitination and subsequent degradation by the proteasome. This Pim degradation presumably leads to activation of Sse. In contrast, Thr does not seem to contain a functional D-box, and mitotic degradation of Thr is dependent on Sse. The initial Thr cleavage event is followed by degradation of the C-terminal cleavage product. Furthermore, rather than activating Sse as in the case of Pim degradation, Thr cleavage contributes to inactivation of Sse (Herzig, 2002).

According to this proposal, degradation of Pim should precede Thr cleavage, since these two events would define a window of Sse activity. Thr cleavage should not occur too fast after Pim degradation so that Sse can cleave its other targets. Thr cleavage therefore might be regulated (for instance, by Scc1 cleavage fragments) or might not lead to Sse inactivation immediately. Sse inactivation might occur only once Thr cleavage fragments have been removed. Alternatively, Sse might cleave its substrates with different kinetics. Fast and efficient Scc1 cleavage may be followed by less efficient and slower Thr cleavage (Herzig, 2002).

It is emphasized that there is no direct evidence for this proposal from biochemical separase activity assays. The assay developed for human separase in the Xenopus extract system does not work for Drosophila Sse complexes for unknown reasons. Perhaps activation of Drosophila Sse complexes is only possible in a particular cellular context, for instance, on the mitotic spindle or at the kinetochore. Consistent with this proposal, only a fraction of Pim and Thr is degraded during mitosis in Drosophila embryos, and a slight enrichment of Pim and Thr on mitotic spindles, similar to securin and separase in yeast, can be visualized with appropriate fixation procedures in the syncytial blastoderm (Herzig, 2002).

Even without biochemical evidence, the data strongly support the notion that Thr is cleaved by Sse. Cleavage occurs at a conserved separase-cleavage consensus sequence. Substitution of a single arginine by an aspartate within this region abolishes cleavage, as previously observed for cleavage of yeast and human Scc1 by separase. Furthermore, mitotic Thr cleavage requires functional Sse complexes, since Thr is neither cleaved in pim mutants, nor in Sse complexes containing nonfunctional Thr mutants, nor in cells arrested in the mitotic checkpoint, when Sse is inactive (Herzig, 2002).

The idea that Thr cleavage and the consequential Thr degradation contribute to inactivation of Sse is supported by genetic analyses. Expression of noncleavable Thr variants results in a phenotype that is highly dependent on the level of Sse protein. The phenotype is only observed with wild-type, but not with reduced levels of Sse. Moreover, noncleavable Thr generates a phenotype only in combination with functional, but not with catalytically inactive Sse, having a serine instead of the cysteine residue in the catalytic center (Herzig, 2002).

Does Thr cleavage represent a general aspect of separase regulation or is it specific for Drosophila? Thr is not conserved during evolution but might correspond to the nonconserved N-terminal domain found in separases from other eukaryotes. Therefore, mitotic Thr cleavage might conceivably correspond to the mitotic separase cleavage, which has been observed in human tissue culture and in vitro. This separase cleavage also appears to be autocatalytic. The cleavage sites in human separase have not yet been mapped precisely, and the functional consequences of cleavage-site mutations are not yet known. However, extrapolating from the reported size of the human separase cleavage fragments to Drosophila, the corresponding processing events should occur within Sse and not within Thr, the putative N-terminal separase domain released during evolution. Sse processing has not been detected in Drosophila. But the hypothesized evolutionary gene split resulting in the independent Sse+ and thr+ genes of Drosophila might represent a permanent separation of those separase fragments that are generated by mitotic cleavage in human cells. The theory that mitotic Thr cleavage does not correspond to human separase self-cleavage is also supported by the apparently distinct functional consequences of these processing events. Whereas Thr cleavage contributes to Sse inactivation, cleaved human separase is clearly active. Mitotic Thr cleavage therefore might be an event specific for insects with their characteristic early embryogenesis including syncytial division cycles followed by cellularization. Early embryogenesis is precisely the developmental period that is most dependent on Thr cleavage. It is not understood why Thr cleavage is essential at 18°C but largely dispensable at 25°C. The reason for this cold-sensitivity is not simply stress per se, because sensitivity was not observed at elevated temperatures. It is noted that microtubule-dependent processes tend to be sensitive to cold temperatures (Herzig, 2002).

At present, it is also not understood why Thr cleavage is particularly crucial for the process of cellularization, whereas it is less important during other developmental stages. As Thr cleavage contributes to Sse inactivation, the phenotypes caused by noncleavable Thr variants presumably reflect Sse hyperactivation. Persistence of Sse activity into S phase might be expected to interfere with the establishment of sister-chromatid cohesion by premature degradation of the Scc1 cohesin subunit. A rapid Sse inactivation resulting from mitotic Thr cleavage, therefore, would be expected to be most important during the extremely rapid syncytial division cycles, during which the alternative pathway of Sse inhibition by resynthesis of the securin Pim during interphase might not be fast enough. In principle, the various irregularities observed during the syncytial cycles in ThrVQ embryos (deleted of the Separase cleavage-site consensus ThrVQ, deletion of amino acids 1031-VEPIRKQ-1037) might reflect consequences from premature Scc1 degradation by hyperactive Sse. The limited penetrance and expressivity of these defects during the syncytial cycles, however, makes a detailed characterization difficult (Herzig, 2002).

The highly penetrant phenotype observed during cellularization is very unlikely to result from premature Scc1 degradation. The extensive cellularization defects start well after completion of mitosis 13, which is at most subtly defective in a few nuclei. It is therefore assumed that hyperactive Sse results in the degradation of an unknown protein that is crucial for cellularization (Herzig, 2002).

Observations in other organisms have also indicated that separase has other targets in addition to cohesin subunits. Caenorhabditis elegans separase appears to have targets whose cleavage is important for osmotic barrier and anterior-posterior axis formation in the fertilized egg. Moreover, a bioinformatics survey has revealed 26 potential separase targets in the S. cerevisiae proteome, and the kinetochore-associated protein Slk19 has in fact been confirmed as a separase target. Cleavage of Slk19 has been shown to contribute to anaphase spindle stability. Even though a Drosophila ortholog for Slk19 cannot be identified, it is conceivable that spindle-associated proteins are also Sse targets in Drosophila. Excess cleavage of microtubule-associated targets important for cytoskeletal organization might thus cause the cellularization defects in Thr DeltaVQ embryos. These embryos clearly have an abnormal gamma-tubulin distribution during interphase 14. The putative additional Sse targets might be exclusively or particularly important during cellularization. Alternatively, it is not excluded that the alternative pathway of Sse inhibition by Pim resynthesis is particularly inefficient before cellularization, because the decrease of maternal pim mRNA levels at this stage might not yet be fully compensated by zygotic pim expression (Herzig, 2002).

In conclusion, although mitotic and meiotic cohesin subunits have been shown to be crucial targets of eukaryotic separases, recent results point to additional substrates involved in processes beyond sister-chromatid separation and to novel regulatory mechanisms. Analyses in different organisms, that have revealed surprisingly distinct aspects of separase regulation and function, will perhaps rapidly converge toward a complete picture (Herzig, 2002).

The cohesin subunit Rad21 is required for synaptonemal complex maintenance, but not sister chromatid cohesion, during Drosophila female meiosis

Replicated sister chromatids are held in close association from the time of their synthesis until their separation during the next mitosis. This association is mediated by the ring-shaped cohesin complex that appears to embrace the sister chromatids. Upon proteolytic cleavage of the alpha-kleisin cohesin subunit at the metaphase-to-anaphase transition by separase, sister chromatids are separated and segregated onto the daughter nuclei. The more complex segregation of chromosomes during meiosis is thought to depend on the replacement of the mitotic alpha-kleisin cohesin subunit Rad21/Scc1/Mcd1 by the meiotic paralog Rec8. In Drosophila, however, no clear Rec8 homolog has been identified so far. Therefore, this study has analyzed the role of the mitotic Drosophila alpha-kleisin Rad21 during female meiosis. Inactivation of an engineered Rad21 variant by premature, ectopic cleavage during oogenesis results not only in loss of cohesin from meiotic chromatin, but also in precocious disassembly of the synaptonemal complex (SC). The lateral SC component C(2)M can interact directly with Rad21, potentially explaining why Rad21 is required for SC maintenance. Intriguingly, the experimentally induced premature Rad21 elimination, as well as the expression of a Rad21 variant with destroyed separase consensus cleavage sites, do not interfere with chromosome segregation during meiosis, while successful mitotic divisions are completely prevented. Thus, chromatid cohesion during female meiosis does not depend on Rad21-containing cohesin (Urban, 2014 -- PubMed ID: 25101996).

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

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

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