InteractiveFly: GeneBrief

Rad21: Biological Overview | References

Every time a cell divides, it is essential that both daughters receive the complete genetic information of their mother. In eukaryotes, the tight connection between the two copies of every chromosome generated by DNA replication (the sister chromatids) ensures the attachment of their kinetochores to spindle microtubules emanating from opposite poles so that the sister chromatids will be segregated into different daughter cells. Cohesin is a highly conserved multisubunit complex that holds sister chromatids together in mitotic cells (see Proposed functions for sister chromatid cohesion proteins in Dorsett, 2007; see A Ring for Holding Sister Chromatids Together?). At the metaphase to anaphase transition, proteolytic cleavage of the α kleisin subunit (Rad21) by separase causes cohesin's dissociation from chromosomes and triggers sister-chromatid disjunction. To investigate cohesin's function in postmitotic cells, where it is widely expressed, fruit flies were created whose Rad21 can be cleaved by TEV protease. Cleavage causes precocious separation of sister chromatids and massive chromosome missegregation in proliferating cells, but not disaggregation of polytene chromosomes in salivary glands. Crucially, cleavage in postmitotic neurons is lethal. In mushroom-body neurons, it causes defects in axon pruning, whereas in cholinergic neurons it causes highly abnormal larval locomotion. These data demonstrate essential roles for cohesin in nondividing cells and also introduce a powerful tool by which to investigate protein function in metazoa (Pauli, 2008).

The investigation of nonmitotic functions of proteins essential for cell proliferation poses a major technical challenge: namely, how to inactivate such proteins without compromising cell proliferation. A good example is the highly conserved multisubunit complex called cohesin, which holds the products of DNA replication (sister chromatids) together and thereby ensures their segregation to opposite poles of the cells during mitosis and meiosis (reviewed in Nasmyth, 2005; Hirano, 2006). Cohesin forms a large tripartite ring (see Schematic of the cohesin complex containing TEV-cleavable Rad21) composed of a pair of Structural Maintenance of Chromosome (SMC) proteins, SMC1 and SMC3, and an α kleisin protein, Scc1/Rad21, whose cleavage by separase causes cohesin's dissociation from chromosomes and triggers sister-chromatid disjunction at the metaphase to anaphase transition. Sister-chromatid cohesion requires two other non-SMC subunits, namely, Pds5 and Scc3/SA, that bind to cohesin's α kleisin subunit. The establishment of cohesion depends on the cohesin loading complex Scc2/Scc4 and on the acetyl-transferase Eco1/Ctf7 (Pauli, 2008).

The fact that cohesin forms a ring whose opening releases it from chromatin has led to the suggestion that it holds sister DNAs together by using a topological mechanism (Gruber, 2003). Importantly, this type of function could also be of value in regulating aspects of chromosome organization that are independent of sister-chromatid cohesion and are not directly required for chromosome segregation. It is notable in this regard that the majority of cohesin is removed from chromosome arms during prophase/prometaphase in most eukaryotic cells by a separase-independent mechanism (Gandhi, 2006; Kueng, 2006). Only cohesin that subsequently persists on chromosomes is cleaved by separase at the onset of anaphase (Waizenegger, 2000). As a consequence, there exists a large pool of cohesin ready to reassociate with chromosomes as soon as cells exit from mitosis during telophase. Cohesin is therefore tightly associated with chromosomes for much of the cell-division cycle and could have important functions on unreplicated genomes (Pauli, 2008).

Much evidence has emerged recently that cohesin might have important roles in regulating gene expression (reviewed in Dorsett, 2007). Approximately half of the cases of a multisystem developmental disorder in humans called Cornelia de Lange syndrome (CdLS), which is characterized by mental retardation, upper limb abnormalities, growth delay, and facial dysmorphisms, are caused by mutations in genes encoding NIPBL/Delangin (the human Scc2 ortholog), SMC1A, or SMC3 (Deardorff, 2007; Krantz, 2004; Musio, 2006; Tonkin, 2004). Because even severe cases of CdLS appear not to be accompanied by defects in sister-chromatid cohesion, it has been suggested that CdLS is caused by misregulated gene expression during embryonic development. Consistent with this possibility, the Drosophila Scc2 ortholog, Nipped-B, facilitates long-range enhancer-promotor interactions, at least for certain genes whose regulatory sequences have been mutated (Dorsett, 2005; Rollins, 1999). Furthermore, mutations in mau-2, the Caenorhabditis elegans Scc4 ortholog, cause defects in axon guidance (Bernard, 2006; Takagi, 1997). Recently, two cohesin subunits, Scc1/Rad21 and SMC3 (Horsfield, 2007), have been implicated in expression of the hematopoietic transcription factors runx1 and runx3 in zebrafish (Pauli, 2008).

Despite these findings, it cannot be excluded that developmental 'cohesinopathies' are in fact caused by 'knock on' effects of compromising the establishment or maintenance of sister-chromatid cohesion. In the case of CdLS, for example, haploinsufficiency of NIPBL/Delangin might cause cell-type-specific sister-chromatid cohesion defects (Kaur, 2005) that would be overlooked by examining this process in only one type of cell. It is therefore vital to develop methods that permit observation of the effects on gene expression and development of eliminating cohesin's function completely without interfering with cell proliferation (Pauli, 2008).

To analyze cohesin's function in a more sophisticated manner than hitherto possible in metazoa, the tobacco etch mosaic virus (TEV) protease was used to cleave cohesin's α kleisin subunit in Drosophila in a cell-type-specific and/or temporally controlled manner. This process opens the cohesin ring and presumably abolishes its topological embrace of chromatin fibers (Gruber, 2003). As expected, expression of TEV protease in proliferating cells of fly embryos whose sole form of Rad21 contains TEV-cleavage sites causes precocious separation of sister chromatids and has a devastating effect on chromosome segregation. More remarkably, TEV-induced Rad21 cleavage in postmitotic neurons is lethal. It causes defects in the developmental axon pruning of mushroom-body γ neurons within pupal brains and defects in cholinergic neurons that result in highly abnormal larval locomotion (Pauli, 2008).

To inactivate cohesin, cleavage of its α kleisin subunit (Rad21) was performed. Although this does not directly affect any known functional domain of Rad21, it severs and thereby opens cohesin's tripartite ring (see Schematic of the cohesin complex containing TEV-cleavable Rad21), leading to its rapid dissociation from chromosomes. To do this in Drosophila, it was necessary (1) to create a Rad21 mutant strain, (2) to complement the Rad21 mutation with a version of Rad21 that contains cleavage sites for a site-specific protease, and (3) to express a version of the protease that can accumulate within nuclei in a tissue-specific and/or time-dependent manner. TEV protease because it has been used successfully for this purpose in the budding yeast Saccharomyces cerevisiae (Pauli, 2008).

The Rad21 gene (CG17436) is located within the centric heterochromatin of chromosome 3L (Markov, 2003), but no mutants were available. To create Rad21 mutations, a P element inserted 4 kb upstream of the transcriptional start of Rad21 was remobilized by P element Transposase. Among the homozygous lethal stocks, four independent Rad21 deletion alleles were identified by using PCR (Rad21^ex3, Rad21^ex8, Rad21^ex15, Rad21^ex16). All four alleles lack exons 1 and 2, which encode the highly conserved N terminus of Rad21 that interacts with the ATPase head of SMC3 (Pauli, 2008).

Although it was known that TEV protease can inactivate protein function in budding yeast, it was unclear whether TEV could be used in a complex metazoan organism. This work shows that TEV can be expressed in a wide variety of Drosophila tissues without causing overt toxicity. More important, TEV expression induces quantitative cleavage of TEV-site-containing, but not wild-type, Rad21 protein; this is accompanied by penetrant phenotypes both in proliferating tissues and, more unexpectedly, in cells not engaged in mitosis, such as neurons and salivary gland cells (Pauli, 2008).

The system has many attractive features that should make it a powerful and versatile tool for studying protein function in vivo. (1) The method causes protein inactivation within a few hours and does not rely on a gradual depletion of the protein, as occurs in methods that interfere with the protein's synthetic capacity, such as recombinase-mediated gene deletion or RNA interference. (2) The system is reversible. By using Gal80ts, TEV protease can be turned both on and off. (3) It is possible to be certain that phenotypes are caused by cleavage of the target protein by comparing the effect of TEV expression in animals whose target protein either does or does not contain TEV sites. (4) By targeting the protease to particular locations inside or even (by using a secreted protease) outside cells, it should be possible to direct inactivation of the target protein to specific intra- or extracellular compartments. The restriction of protein inactivation to specific cellular compartments may be easier to devise by using TEV than degron systems relying on the much more complex process of ubiquitin-mediated proteolysis. Unlike the MARCM system, which uses FLP/FRT-induced mitotic recombination to generate homozygous mutant clones in proliferating tissues, TEV cleavage can be triggered in all cells of a given tissue and at any stage of development, features that will greatly facilitate phenotypic and biochemical analyses. Because many eukaryotic proteins contain multiple functional domains connected by unstructured polypeptide chains, protein inactivation through TEV cleavage should be applicable to a large variety of proteins. It could also be used to clip off protein domains and thereby alter protein activity (Pauli, 2008).

The first priority upon developing a system to cleave Rad21 was to use it to investigate the role of cohesin during mitosis. In yeast, cohesin has a vital role in holding sister chromatids together until all chromosomes have bioriented during mitosis, whereupon cleavage of Scc1/Rad21 by separase triggers sister-chromatid disjunction (reviewed in Nasmyth, 2005). The consequences of depleting Scc1/Rad21 from tissue-culture cells by using RNA interference are, on the whole, consistent with the above-mentioned notion (Coelho, 2003; Vass, 2003). However, results from depletion experiments have not been able to directly explain the effects of inactivating cohesin within a single cell cycle (Pauli, 2008).

A situation was engineered in which efficient cleavage of Rad21 occurred precisely as embryonic cells embarked on cycle 14, causing a devastating effect on mitosis. Chromosomes enter mitosis with paired sister kinetochores; however, instead of stably biorienting on a metaphase plate, they disjoin precociously, usually segregating to opposite poles. Importantly, these highly abnormal movements all take place prior to the APC/C-dependent activation of separase. These observations imply that cohesin is essential for the sister-chromatid cohesion necessary to resist mitotic-spindle forces in metazoan organisms as well as in yeast (Pauli, 2008).

The finding that most sister chromatids (in cells with cleaved Rad21) disjoin to opposite spindle poles, albeit precociously, suggests that their chromosomes possess sufficient cohesion to establish a transient form of biorientation, though possibly with low accuracy. At this stage it is not possible to determine whether this cohesion is mediated by cohesin complexes that have survived Rad21 TEV cleavage or by an independent cohesive mechanism such as residual sister DNA catenation. It can nevertheless be concluded that the latter, if it exists, is incapable of resisting spindle forces and cannot therefore maintain sister-chromatid cohesion during a period in which the spindle-assembly checkpoint (SAC) has been activated and errors in chromosome biorientation are corrected. Thus, what really distinguishes cohesion mediated by cohesin from DNA catenation is its ability to be regulated by the SAC, and this may be the reason why eukaryotic cells appear to use cohesin for mitosis (Pauli, 2008).

Mutations in Scc2's human ortholog as well as in SMC1 and SMC3 cause the developmental defects associated with CdLS (reviewed in Dorsett, 2007). It is unclear whether these defects are caused by mitotic errors during development or by defects in nonmitotic cohesin functions. The first clue that cohesin might indeed play key roles during development other than holding sister chromatids together was the finding that mutations in Drosophila Nipped-B (Rollins, 1999), the ortholog of Scc2, alters the expression of genes whose regulatory sequences have been mutated (Pauli, 2008).

If cohesin has nonmitotic functions during development, then these could occur in proliferating and nonproliferating (postmitotic) cells. To analyze cycling cells, it would be necessary to restrict analysis either to a short, specific cell-cycle stage (e.g., the G1 period) or to develop a means of differentially inactivating cohesin complexes engaged in nonmitotic functions, leaving intact those engaged in chromosome segregation. Analysis of postmitotic cells is easier. It is merely necessary to devise a protocol for inactivating cohesin only after cell proliferation has ceased (Pauli, 2008).

Cleavage of Rad21 induced by postmitotic pan-neuronal drivers causes lethality, suggesting that cohesin has key functions in neurons. To investigate these in greater detail, the effects of Rad21 cleavage in specific neuronal subtypes were analyzed. The finding that the proliferative defects caused by a SMC1 mutation in clones of mushroom-body neuroblasts are accompanied by defective pruning of axons (Schuldiner, 2008) led to an investigation of the effects of Rad21 cleavage in γ neurons. The results show that Rad21 cleavage abolishes the developmentally controlled pruning of both axons and dendrites in γ neurons. These defects cannot have been caused by failures in cell division because cleavage had no effect on the birth of γ neurons or on their initial axonal projections (Pauli, 2008).

Previous work on mau-2 (the C. elegans Scc4 ortholog) has already provided a link between cohesin and axon development (Benard, 2004). Whereas Mau-2 was reported to act as a guidance factor required for correct axon and cell migration, investigation of γ neurons in Drosophila suggests that cohesin mediates the elimination of axon projections and dendrites. However, the results do not rule out a function for cohesin in regulating axon guidance because Rad21 cleavage might not be complete when γ-neuron axons start growing out in the first place. Indeed, axon-projection defects were detected in developmentally arrested late pupae (Pauli, 2008).

It has not thus far been possible to show that γ-neuron pruning defects cause changes in animal behavior. Cleavage of cohesin in the entire population of cholinergic neurons, in contrast, has a dramatic effect, causing larvae to turn frequently, move their heads back and forth, and even crawl backward. Importantly, the neurons clearly survive without functional cohesin and must be at least partially active, because larvae are not paralyzed by cohesin cleavage, a phenotype seen when cholinergic transmission is switched off. The locomotion defects are not dissimilar to those caused by mutations in scribbler (sbb). sbb, also known as brakeless (bks) and master of thickveins (mtv), codes for a ubiquitously expressed corepressor of transcription (Haecker, 2007; references therein). Expression of a sbb transcript exclusively in cholinergic neurons is sufficient to rescue locomotion defects of sbb mutants (Suster, 2004). It therefore appears that the lack of sbb and cohesin in cholinergic neurons causes similar locomotion defects. Future work will have to show whether there is a link between sbb and cohesin. The finding that cohesin has roles in neurons that are essential for normal behavior is consistent with the notion that the mental retardation invariably found in patients with CdLS is also due to defective neuronal function, as opposed to defective cell proliferation during development (Pauli, 2008).

This study has shown that suppression of 201Y-Gal4-induced TEV expression, specifically in muscles, bypasses the early pupal arrest in Rad21^TEV-rescued flies and indicates that cohesin is essential in muscles as well as in neurons. In addition, although cohesin does not seem to be required for the maintenance of polytene-chromosome morphology, it is essential for normal progression through the endocycle in salivary glands. It is therefore conceivable that cohesin has key functions in most postmitotic cell types. What might these functions be? Cohesin is known to be required for efficient double-strand break repair as well as sister-chromatid cohesion (reviewed in Nasmyth, 2005), and it promotes repair by facilitating homologous recombination between sister chromatids. Its action in postmitotic neurons, however, must be on unreplicated chromatids. It is suggested therefore that cohesin's function in neurons and other postmitotic G0 cells is more likely to be in regulating gene expression. The finding that cohesin cleavage reduces the accumulation of EcR-B1 within γ neurons is consistent with this notion. Interestingly, recent data have shown that cohesin binds to the EcR gene in several fly cell lines (Misulovin, 2009). Future experiments should address whether cohesin acts as a general regulator of gene expression (Pauli, 2008).

In summary, definitive evidence is provided that the cohesin ring has essential functions in cells with unreplicated chromosomes. It will be important in the future to establish whether cohesin functions by trapping chromatin fibers, as it appears to do in cells that have replicated their genomes (Pauli, 2008).

Chromosomal cohesin forms a ring

The cohesin complex is essential for sister chromatid cohesion during mitosis. Its Smc1 and Smc3 subunits are rod-shaped molecules with globular ABC-like ATPases at one end and dimerization domains at the other connected by long coiled coils. Smc1 and Smc3 associate to form V-shaped heterodimers. Their ATPase heads are thought to be bridged by a third subunit, Scc1, creating a huge triangular ring that could trap sister DNA molecules. This study addressed whether cohesin forms such rings in vivo. Proteolytic cleavage of Scc1 by separase at the onset of anaphase triggers its dissociation from chromosomes. N- and C-terminal Scc1 cleavage fragments remain connected due to their association with different heads of a single Smc1/Smc3 heterodimer. Cleavage of the Smc3 coiled coil is sufficient to trigger cohesin release from chromosomes and loss of sister cohesion, consistent with a topological association with chromatin (Gruber, 2003; full text of article).

The preferential association of Scc1's N-terminal fragment with Smc3's head and its C-terminal one with that of Smc1 is conserved in its meiotic counterpart Rec8. Scc1 and Rec8 share very little sequence homology except within their first and last 100 amino acids, which are generally conserved amongst Scc1 and Rec8 homologs from a wide variety of eukaryotes. These conserved terminal sequences must contain the SMC head interaction domains because fragments containing the first 115 amino or the last 115 amino acids of Scc1 formed complexes specifically with Smc3 or Smc1, respectively. There is no apparent similarity between Scc1's Smc3 binding N-terminal 115 amino acids with its Smc1 binding 115 C-terminal ones. This asymmetry presumably corresponds to an asymmetry of the Smc1/3 heterodimer's two heads, to which these sequences bind. The finding that these N- and C-terminal domains of Scc1 and Rec8 are homologous to those of ScpA proteins in bacteria and barren subunits of condensin implies that the heads of all SMC-like proteins may be connected in a similar manner. It is for this reason that that members of this family are called kleisins (from the Greek word for closure: kleisimo). Cohesin's asymmetry is presumably shared by condensin, which contains an Smc2/Smc4 heterodimer but contrasts with the symmetry of bacterial SMC proteins. If the latter bind the ScpA, as is currently suspected, then the molecular symmetry of the SMC dimer would predict that they bind at least two molecules of ScpA in a symmetric fashion (Gruber, 2003).

This image of Scc1-SMC interactions, inferred merely from co-expression of different subunits and their fragments in insect cells, is in full agreement with analysis of cohesin released from yeast chromosomes after cleavage by separase in vivo. Thus, both Scc1's N-terminal cleavage fragment and a fragment containing Smc3's head domain are released from the rest of separase-cleaved cohesin by severance of Smc3's coiled coil, while the C-terminal Scc1 fragment remains associated with the central domain severed from its heads, presumably due to the latter's association with Smc1. Soluble yeast cohesin appears to possess the same fundamental geometry; namely Scc1's N-terminal half is connected with Smc3's head while Scc1's C-terminal half is connected to Smc3's central domain via its interaction with Smc1 (Gruber, 2003).

The ring hypothesis postulates that Scc1, Smc1 and Smc3 form a triangle and predicts that all three subunits and fragments therefore should remain associated with each other when any single side of the triangle has been broken irrespective of where this break is. The finding that Smc3 holds together N- and C-terminal cleavage fragments of Scc1 while Scc1 holds together the two halves of Smc3 when its arm has been severed by double cleavage with the TEV protease clearly confirms this prediction. It is remarkable that breakage of any one the triangle's sides destroys its closure and hence its potential for trapping DNA without affecting its three corners, i.e., subunit interactions (Gruber, 2003).

Though these findings together with the electron micrographs of soluble cohesin are consistent with cohesin being an asymmetric monomeric complex, they do not by themselves rule out the possibility that it forms dimers when bound to chromatin, especially after DNA replication when cohesin complexes on different sister chromatids could in principle interact to generate cohesion. Available evidence is, however, inconsistent with this notion. For example, differently tagged versions of the same cohesin subunit cannot be coimmunoprecipitated from cohesin released from chromatin by micrococcal nuclease digestion. One specific dimer model would have it that the two Scc1 molecules within a cohesin dimer connect Smc1 and Smc3 heads not from the same but from different heterodimers. If this were so, then one might predict that cleavage of just one of the Scc1 molecules might be sufficient to release cohesin dimers from chromatin. This study found that TEV protease treatment of chromatin from a heterozygous diploid strain carrying wild-type and TEV cleavable Scc1 released Scc1 fragments but not intact Scc1. Therefore the notion is favored that both soluble and chromatin bound cohesin form an asymmetric monomeric complex (Gruber, 2003).

An important but yet unresolved issue is the mechanism by which the cohesin and other SMC protein-containing complexes bind to chromosomes. One model predicts that each of the two SMC head domains binds to one sister chromatid. It has been suggested on the basis of electron spectroscopic images of condensin complexes bound to DNA that the double helix might be wrapped around the SMC head domains. It is, however, unclear from the existing crystal structure of these domains how DNA could conceivably be wrapped round a domain whose dimensions are much smaller than those of a nucleosome (Gruber, 2003).

The finding that chromosomal cohesin forms a closed ring raises the possibility that cohesin's interaction with DNA is topological and not chemical in nature. By passing through cohesin's ring, chromosomal DNA could be topologically trapped by cohesin. This model makes no prediction as to the actual path of DNA as it passes through cohesin's ring and does not exclude the possibility of DNA being wrapped around the ring as opposed to simply passing once through it. A topological interaction of this nature would explain the resistance to high salt concentrations of cohesin's association with chromosomes as well as its resistance to DNA intercalating agents such as ethidium bromide. It would also explain why the ring's severance, be it within Scc1 or Smc3, is sufficient to release cohesin from chromosomes either in vitro or in vivo without apparently affecting any of cohesin's subunit interactions. If cohesin embraces chromosomal DNA in this manner, then sister chromatid cohesion could arise from the passage of sister DNA molecules through the same cohesin ring. Sister DNA molecules could also conceivably pass through different but interacting cohesin rings; though this would predict hitherto undetected interactions between cohesin complexes. Further experiments will be required to establish whether DNA really passes inside cohesin's ring (Gruber, 2003).

Though the experiments are consistent with DNA's passage through cohesin's ring, they shed little or no insight into the mechanism by which transient opening of the ring permits entry of the double helix. Such a process would be analogous to the entry of a climbing rope into a carabiner, which is a ring with a gate. In this model, Scc1 could be considered cohesin's gate and the binding of ATP to the SMC head domains and/or its hydrolysis might regulate the opening and shutting of cohesin's gate. A key question is whether Scc1 connects the head domains of Smc1 and Smc3 when they are bound to ATP and have thereby themselves 'dimerized' or whether it only connects the two heads after ATP has been hydrolyzed. In the latter case, cohesin could switch between two types of ring: one in which the heads are bound to each other but the Scc1 gate is open and another in which Scc1 alone connects the heads. Such a system would permit DNA strands to enter cohesin's ring without pre-existing strands exiting. Even if this hypothesis is correct, it remains a mystery how some gates can apparently be reopened without destroying Scc1, as presumably occurs during prophase in metazoan cells, while others can only be opened by cleavage of Scc1 at anaphase onset (Gruber, 2003).

Given the similar structure of Scc1, barren, ScpA, and the other members of the kleisin superfamily, it is hard not to believe that all complexes composed of SMC proteins operate using a similar topological principle to that proposed for cohesin, namely passage of DNA inside two arms held together by Scc1-like bridges (kleisins), which can be opened and shut. Such devices are presumably indispensable for regulating the packing of DNA within cells because unlike nucleosomes, they clearly existed in the common ancestor of all life forms on this planet and have been retained in almost all organisms ever since. Cohesin may be unique in holding sister DNAs within a single ring whereas condensin may have the ability to pass the same DNA molecule more than once through its ring, thereby acting as a DNA coil securing device (Gruber, 2003).

Drosophila Nipped-B protein supports sister chromatid cohesion and opposes the Stromalin/Scc3 cohesion factor to facilitate long-range activation of the cut gene

The Drosophila melanogaster Nipped-B protein facilitates transcriptional activation of the cut and Ultrabithorax genes by remote enhancers. Sequence homologues of Nipped-B, Scc2 of Saccharomyces cerevisiae, and Mis4 of Schizosaccharomyces pombe are required for sister chromatid cohesion during mitosis. The evolutionarily conserved Cohesin protein complex mediates sister chromatid cohesion, and Scc2 and Mis4 are needed for Cohesin to associate with chromosomes. This study shows that Nipped-B is also required for sister chromatid cohesion but that, opposite the effect of Nipped-B, the stromalin/Scc3 component of Cohesin inhibits long-range activation of cut. To explain these findings, a model is proposed based on the chromatin domain boundary activities of Cohesin in which Nipped-B facilitates cut activation by alleviating Cohesin-mediated blocking of enhancer-promoter communication (Rollins, 2004).

These experiments addressed two questions: (1) does Nipped-B, in addition to facilitating remote activation of cut and Ultrabithorax, participate in mitotic sister chromatid cohesion, and (2) does Cohesin participate in long-range activation of cut? The first was motivated by the sequence similarity of Nipped-B to yeast adherins required for sister chromatid cohesion, and the second was motivated by the published observations that the yeast adherins are required for the Cohesin complex to associate with chromosomes. The results indicate that the cooperation between adherins and Cohesin that occurs in yeast is conserved in Drosophila. The findings also indicate, however, that the SA/Scc3 subunit of Cohesin opposes Nipped-B in long-range activation of the cut gene (Rollins, 2004).

The high rate of precocious sister chromatid separation (PSCS) in homozygous and heteroallelic Nipped-B mutants and the increased lethality of Rad21/Scc1 RNAi in flies heterozygous for a Nipped-B mutation observed in this study are consistent with the findings that the yeast Scc2 and Mis4 homologues of Nipped-B are required for Cohesin to associate with chromosomes. Although PSCS was detected in third-instar neuroblasts in RNAi experiments, it is unlikely that the synthetic lethality of a heterozygous Nipped-B mutation and Rad21/Scc1 RNAi is caused by changes in long range gene activation, because as discussed below, Nipped-B and Cohesin appear to have opposing roles in gene activation. In the RNAi experiments, it is possible that the pupal lethality is caused by PSCS in a subpopulation of critical cells, but the possibility cannot be ruled out that Nipped-B and Cohesin cooperate with each other in other essential functions (Rollins, 2004).

No physical association between the yeast adherins and Cohesin has been detected, nor do they colocalize on chromosomes. No chromosomal population of Nipped-B could be detected, but if Nipped-B acts primarily as a chaperone for loading Cohesin onto chromosomes, only a small fraction of Nipped-B may be transiently interacting with chromosomes at any time. The mechanisms by which yeast adherins facilitate Cohesin chromosome binding are unclear, but the synthetic lethality between a heterozygous Nipped-B mutation and Rad21/Scc1 RNAi observed here indicates that this functional connection is conserved in metazoans. In budding yeast, Cohesin begins to associate with chromosomes in late G₁, while in fission yeast, Caenorhabditis elegans, Drosophila, and mammalian cells it begins to associate in telophase. Nipped-B is detected in the nucleus at all stages that have a nuclear membrane, indicating that it could be involved in Cohesin chromosomal association beginning in telophase and thus could influence all potential interphase functions of Cohesin, in addition to sister chromatid cohesion (Rollins, 2004).

The finding that SA/Scc3 RNAi reduces the severity of the ct^K wing-nicking phenotype indicates that the SA/Scc3 component of Cohesin inhibits cut expression. This is the opposite to the role of Nipped-B at cut. Multiple Nipped-B mutations were recovered in a screen for mutations that increase the severity of a wing-nicking phenotype displayed by a cut allele with a weak gypsy insulator insertion. Reduced mRNA levels indicated that some of these Nipped-B mutations are loss-of-function alleles, and viability of homozygous Nipped-B mutants was rescued by a transgene expressing a Nipped-B cDNA from a Chip gene promoter. Thus, Nipped-B protein facilitates activation of cut by the wing margin enhancer (Rollins, 2004).

The effect of Nipped-B on cut expression is likely direct. Nipped-B does not regulate cut by altering the activities of known cut regulators because it is most limiting for cut expression when there is a gypsy insertion at cut while the other known regulators are more limiting with other types of cut mutations. Moreover, heterozygous Nipped-B loss-of-function alleles reduce cut expression, and partial reduction of Nipped-B is unlikely to cause an equal or greater change in the expression of another cut regulator. Although the effects of Nipped-B on gene expression were most apparent with gypsy insertion alleles of cut, a measurable effect was observed in heterozygous females with a wild-type cut allele and an allele in which the wing margin enhancer is deleted. Thus, Nipped-B also facilitates the activation of wild-type cut (Rollins, 2004).

All three SA/Scc3 RNAi insertions and one of three Rad21/Scc1 insertions reduced the number of nicks displayed by the ct^K gypsy insertion. It is thought likely that the Cohesin complex, and not just one or two of its subunits, is responsible for reducing cut expression. Scc1 and Scc3 operate together as a unit in both Drosophila and C. elegans. Thus, it is unlikely that they work independently of each other in regulating gene expression. Indeed, Rad21/Scc1 RNAi in cultured Drosophila cells reduces both Rad21/Scc1 and SA/Scc3 proteins, and data presented here indicate that Rad21/Scc1 and SA/Scc3 may regulate each other's transcript levels. However, the possibility cannot be ruled out that Rad21/Scc1 and SA/Scc3 work independently of the Smc1 and Cap/Smc3 Cohesin subunits, which form another stable subcomplex. Initial attempts to reduce expression of the SMC subunits by RNAi were unsuccessful (Rollins, 2004).

It is unlikely that the effects of SA/Scc3 on cut expression occur by reducing the expression of a cut activator. The small reductions in SA/Scc3 expression in these experiments are unlikely to cause equal or larger changes in the activities of other cut regulators. Also effects are not seen of Cohesin RNAi on ct^53d, which has a small deletion in the enhancer and is affected by all known cut regulators except Nipped-B. It is most likely, therefore, that SA/Scc3 acts directly at cut or by reducing the ability of Nipped-B to facilitate activation (Rollins, 2004).

The possibility that the negative effect of SA/Scc3 on cut expression may be specific to gypsy insertion alleles cannot be ruled out, it is thought improbable. The negative effect is likely to be related to the positive effect of Nipped-B, and Nipped-B facilitates the expression of wild-type cut. If SA/Scc3 does specifically affect gypsy insertion alleles, however, it may interact with the gypsy insulator and contribute to enhancer blocking. This is consistent with evidence that Cohesin functions at chromosomal boundaries in yeast. Certain Smc1 and Smc3 mutations reduce the ability of a boundary that flanks the HMR silent mating-type locus to block the spread of gene-silencing Sir protein complexes, and Scc1 associates with this boundary. It has also been proposed that Cohesin binding sites are boundaries that control the extent of chromosome loop formation by Condensin. This proposal is based in part on the observation that Cohesin is needed to reestablish chromosome condensation upon returning temperature-sensitive Condensin mutants to the permissive temperature. In Drosophila, the gypsy insulator partially blocks the negative effects of heterochromatin on the expression of a euchromatic gene, suggesting that it has boundary activity, and in yeast, the Su(Hw) protein that binds the gypsy insulator also blocks the spread of gene-silencing complexes. If SA/Scc3 or Cohesin increases insulation by gypsy, Nipped-B could facilitate activation by reducing their association with the insulator (Rollins, 2004).

A more general version of the 'Cohesin insulator' model is preferred, in which native Cohesin binding sites in the 85-kb region separating the wing margin enhancer from the cut promoter act as insulators and impede the formation of structures needed to bring the wing margin enhancer close to the promoter. In yeast, Cohesin binds every 10 kb or so along the chromosomes. The spacing of Cohesin in Drosophila has not been investigated, but multiple complexes could bind in the 85-kb interval between the wing margin enhancer and the cut promoter. Assuming that Nipped-B, perhaps by opening the Cohesin ring, facilitates both the loading and the removal of Cohesin from chromosomes, could explain how Nipped-B facilitates the activation of wild-type cut. By opening the Cohesin ring, Nipped-B would help achieve equilibrium between the bound and unbound states by providing opportunities to load or remove Cohesin from chromosomes. This mechanism would be distinct from proteolytic removal of Cohesin by separase at the metaphase-to-anaphase transition but could be involved in the removal of Cohesin from chromosome arms in prophase. In heterozygous Nipped-B mutants, which retain substantial Nipped-B activity, the equilibrium endpoint would not be altered, but it might take longer to achieve equilibrium. Thus, reduced Cohesin binding to chromosomes would not be expected, but the lower Nipped-B levels would reduce the windows of opportunity for removal of Cohesin needed to allow long-range activation. This model also predicts that Nipped-B does not have to stably associate with chromosomes, which could explain why no chromosomally bound Nipped-B was detected by immunostaining (Rollins, 2004).

Finally, in a simple indirect model it could be supposed that, similar to its role in loading Cohesin, Nipped-B could also facilitate chromosomal binding of another protein complex that assists long-range enhancer-promoter interactions. In this case, there would be competition between Cohesin and the long-range activation complex for Nipped-B, and reduction of Cohesin would make Nipped-B more available to facilitate long-range activation. This and the insulator model described above are not mutually exclusive, but both explain how Nipped-B cooperates with Cohesin in sister chromatid cohesion but opposes the effect of Cohesin proteins on cut expression (Rollins, 2004).

Depletion of Drad21/Scc1 in Drosophila cells leads to instability of the cohesin complex and disruption of mitotic progression

The coordination of cell cycle events is necessary to ensure the proper duplication and dissemination of the genome. The consequences of depleting Rad21 and SA (also known as Scc3; Losada, 1998), two non-SMC subunits of the cohesin complex, by dsRNA-mediated interference in Drosophila cultured cells, were examined. A bona fide cohesin complex exists in Drosophila embryos. Strikingly, the Drad21/Scc1 and SA/Scc3 (Stromalin) non-SMC subunits associate more intimately with one another than they do with the SMCs. Defects were observed in mitotic progression in cells from which Drad21 has been depleted: cells delay in prometaphase with normally condensed, but prematurely separated, sister chromatids and with abnormal spindle morphology. Much milder defects are observed when SA is depleted from cells. The dynamics of the chromosome passenger protein, INCENP, are affected after Drad21 depletion. The surprising observation was made that SA is unstable in the absence of Drad21; however, the converse is not true. Interference with Drad21 in living Drosophila embryos also has deleterious effects on mitotic progression. It is concluded that Drad21, as a member of a cohesin complex, is required in Drosophila cultured cells and embryos for proper mitotic progression. The protein is required in cultured cells for chromosome cohesion, spindle morphology, dynamics of a chromosome passenger protein, and stability of the cohesin complex, but apparently not for normal chromosome condensation. The observation of SA instability in the absence of Drad21 implies that the expression of cohesin subunits and assembly of the cohesin complex are be tightly regulated (Vass, 2003; Online text).

This study demonstrates that Drad21 is in a cohesin complex with SMC1, SMC3, and SA in Drosophila embryos. Drad21 depletion results in SA instability; intriguingly, however, the converse is not true. This result suggests that SA must interact with Drad21 in order to be stable (perhaps SA is synthesized only after Drad21 accumulates in the cell). This may help to ensure a 1:1 ratio between these subunits (as observed in cohesin complexes in S. cerevisiae; Haering, 2002). Upon Drad21 depletion, dramatic effects are seen on mitotic progression; cells are delayed in prometaphase with prematurely separated sister chromatids and abnormal spindle morphology. In contrast, no premature separation of sister chromatids was seen or significant effects on the cell cycle when SA is depleted, suggesting that the Drad21 phenotype is likely specific to the interference with Drad21 only (Vass, 2003).

Chromosome condensation in either the Drad21- or the SA-depleted cells appeared normal, as judged by overall size and shape of chromosomes, localization of the Barren non-SMC condensin subunit, and the centromeric domain occupied by the CID centromeric protein. Since chromosomes also exhibited normal condensation in Scc1-knockout DT40 cells, human cells expressing a dominant-negative N-terminal truncation of Scc1, and after immunodepletion of cohesin from Xenopus egg extracts, it appears likely that cohesins act independently of condensation machinery in metazoan chromosome structure (Vass, 2003).

The dispersed single chromatids observed in Drad21-depleted Drosophila cells were in contrast to the separate, albeit proximal, chromatids after the knockout of Scc1 from DT40 chicken cells. Perhaps for that reason, INCENP appeared along chromosome arms in metaphase Scc1-knockout DT40 cells (INCENP distribution in later mitotic stages was not reported). In contrast, in early mitotic S2 cells depleted of Drad21, INCENP appeared diffusely localized, possibly because chromatids were no longer close to one another. Later, mitotic cells with aberrant INCENP localization fell into three groups: cells that displayed INCENP staining on single chromatids, cells that showed INCENP transferring onto the spindle even though chromatids had failed to congress to a metaphase plate, and cells in which INCENP associated with microtubules, but was not restricted to the central spindle. Drad21-depleted cells that progressed into the final mitotic stages indicated that INCENP could, however, still localize to the central region of the cell, even though chromosome segregation had not occurred. These results suggest that the correct localization of INCENP to the centromeric domain and its subsequent translocation to the central spindle at the metaphase-to-anaphase transition is dependent on the presence of cohesion between sister centromeres. However, even in the absence of chromatid cohesion, INCENP was able to achieve microtubule localization, albeit in an abnormal temporal and spatial manner (Vass, 2003).

It is unlikely that the separate chromatids observed after Drad21 depletion would form bipolar spindle attachments and align at a metaphase plate. The metaphase checkpoint should be activated, resulting in prometaphase delay. A potential role for sister chromatid cohesion and kinetochore attachment in the metaphase checkpoint has been suggested, with the correct alignment of all sister kinetochores clearly required to establish bipolarity and loss of Mad2 metaphase checkpoint signaling. Why cells are delayed and not arrested by the metaphase checkpoint may be a reflection of compromised checkpoints in Drosophila cultured cells (derived from embryos), since these cells are extremely difficult to synchronize in response to numerous cell cycle inhibitors (Vass, 2003).

The Drosophila RAD21 cohesin persists at the centromere region in mitosis

'Cohesin' is a highly conserved multiprotein complex thought to be the primary effector of sister-chromatid cohesion in all eukaryotes. 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. This study addressed this apparent anomaly by investigating the cell-cycle dynamics of DRAD21, the Drosophila orthologue of the Xenopus XRAD21 and Saccharomyces cerevisiae Scc1p/Mcd1p cohesins. Analysis of DRAD21 in S2 Drosophila tissue culture cells and live embryos expressing a DRAD21-green fluorescent protein (GFP) fusion revealed the presence of four distinct subcellular pools of DRAD21: a cytoplasmic pool; a chromosome-associated pool which dissociates from chromatin as chromosomes condense in prophase; a short-lived centrosome-associated pool present during metaphase-anaphase; and a centromere-proximal pool which remains bound to condensed chromosomes, is found along the junction of sister chromatids between kinetochores, and persists until the metaphase-anaphase transition. It is concluded that in Drosophila, and possibly all metazoans, a minor pool of cohesin remains binds to centromere-proximal chromatin after prophase and maintains sister-chromatid cohesion until the metaphase-anaphase transition (Warren, 2000; full text of article).

Sequential loading of cohesin subunits during the first meiotic prophase of grasshoppers

The cohesin complexes play a key role in chromosome segregation during both mitosis and meiosis. They establish sister chromatid cohesion between duplicating DNA molecules during S-phase, but they also have an important role during postreplicative double-strand break repair in mitosis, as well as during recombination between homologous chromosomes in meiosis. An additional function in meiosis is related to the sister kinetochore cohesion, so they can be pulled by microtubules to the same pole at anaphase I. Data about the dynamics of cohesin subunits during meiosis are scarce; therefore, it is of great interest to characterize how the formation of the cohesin complexes is achieved in order to understand the roles of the different subunits within them. This study investigated the spatio-temporal distribution of three different cohesin subunits in prophase I grasshopper spermatocytes. Structural maintenance of chromosome protein 3 (SMC3) appears as early as preleptotene, and its localization resembles the location of the unsynapsed axial elements, whereas radiation-sensitive mutant 21 (RAD21) (sister chromatid cohesion protein 1, SCC1) and stromal antigen protein 1 (SA1) (sister chromatid cohesion protein 3, SCC3) are not visualized until zygotene, since they are located in the synapsed regions of the bivalents. During pachytene, the distribution of the three cohesin subunits is very similar and all appear along the trajectories of the lateral elements of the autosomal synaptonemal complexes. However, whereas SMC3 also appears over the single and unsynapsed X chromosome, RAD21 and SA1 do not. It is concluded that the loading of SMC3 and the non-SMC subunits, RAD21 and SA1, occurs in different steps throughout prophase I grasshopper meiosis. These results strongly suggest the participation of SMC3 in the initial cohesin axis formation as early as preleptotene, thus contributing to sister chromatid cohesion, with a later association of both RAD21 and SA1 subunits at zygotene to reinforce and stabilize the bivalent structure. Therefore, it is speculated that more than one cohesin complex participates in the sister chromatid cohesion at prophase I (Valdeolmillos, 2007).

Search PubMed for articles about Drosophila Rad21

Benard, C. Y., Kebir, H., Takagi, S. and Hekimi, S. (2004). mau-2 acts cell-autonomously to guide axonal migrations in Caenorhabditis elegans. Development 131(23): 5947-58. PubMed citation: 15539489

Bernard, P., et al. (2006). A screen for cohesion mutants uncovers Ssl3, the fission yeast counterpart of the cohesin loading factor Scc4. Curr. Biol. 16: 875-881. PubMed citation: 16682348

Coelho, P. A., Queiroz-Machado, J. and Sunkel, C. E. (2003). Condensin-dependent localisation of topoisomerase II to an axial chromosomal structure is required for sister chromatid resolution during mitosis. J. Cell Sci. 116: 4763-4776. PubMed citation: 14600262

Deardorff, M. A., et al. (2007). Mutations in cohesin complex members SMC3 and SMC1A cause a mild variant of cornelia de Lange syndrome with predominant mental retardation. Am. J. Hum. Genet. 80: 485-494. PubMed citation: 17273969

Dorsett, D., et al. (2005). Effects of sister chromatid cohesion proteins on cut gene expression during wing development in Drosophila. Development 132(21): 4743-53. PubMed citation: 16207752

Dorsett, D. (2007). Roles of the sister chromatid cohesion apparatus in gene expression, development, and human syndromes. Chromosoma 116(1): 1-13. PubMed citation: 16819604

Gandhi, R., Gillespie, P. J. and Hirano, T. (2006). Human Wapl is a cohesin-binding protein that promotes sister-chromatid resolution in mitotic prophase. Curr. Biol. 16: 2406-2417. PubMed citation: 17112726

Gruber S. Haering, C. H. and Nasmyth, K. K. (2003). Chromosomal cohesin forms a ring. Cell 112: 765-777. PubMed citation: 12654244

Haecker, A., et al. (2007). Drosophila brakeless interacts with atrophin and is required for tailless-mediated transcriptional repression in early embryos. PLoS Biol. 5(6): e145. PubMed citation: 17503969

Haering, C. H. Lowe, J. Hochwagen, A. and Nasmyth, K. (2002). Molecular architecture of SMC proteins and the yeast cohesin complex. Mol. Cell 9: 773-788. PubMed citation: 11983169

Hirano, T. (2006). At the heart of the chromosome: SMC proteins in action. Nat. Rev. Mol. Cell Biol. 7: 311-322. PubMed citation: 16633335

Horsfield, J. A., et al. (2007). Cohesin-dependent regulation of Runx genes. Development 134: 2639-2649. PubMed citation: 17567667

Kaur, M., et al. (2005). Precocious sister chromatid separation (PSCS) in Cornelia de Lange syndrome. Am. J. Med. Genet. A 138: 27-31. PubMed citation: 16100726

Krantz, I. D., et al. (2004). Cornelia de Lange syndrome is caused by mutations in NIPBL, the human homolog of Drosophila melanogaster Nipped-B. Nat. Genet. 36: 631-635. PubMed citation: 15146186

Kueng, S., et al. (2006). Wapl controls the dynamic association of cohesin with chromatin. Cell 127: 955-967. PubMed citation: 17113138

Losada, A., Hirano, M. and Hirano, T. (1998). Identification of Xenopus SMC protein complexes required for sister chromatid cohesion. Genes Dev. 12: 1986-1997. PubMed citation: 9649503

Markov, A. V. et al. (2003). Localization of cohesin complexes of polytene chromosomes of Drosophila melanogaster located on interbands. Genetika 39: 1203-1211. PubMed citation: 14582389

Misulovin, Z., et al. (2008). Association of cohesin and Nipped-B with transcriptionally active regions of the Drosophila melanogaster genome. Chromosoma 117(1): 89-102. PubMed citation: 17965872

Musio, A., et al. (2006). X-linked Cornelia de Lange syndrome owing to SMC1L1 mutations. Nat. Genet. 38: 528-530. PubMed citation: 16604071

Nasmyth, K. and Haering, C. H. (2005). The structure and function of SMC and kleisin complexes. Annu. Rev. Biochem. 74: 595-648. PubMed citation: 15952899

Pauli, A., et al. (2008). Cell-type-specific TEV protease cleavage reveals cohesin functions in Drosophila neurons. Dev. Cell 14(2): 239-51. PubMed citation: 18267092

Rollins, R. A., Morcillo, P. and Dorsett, D. (1999). Nipped-B, a Drosophila homolog of chromosomal adherins, participates in activation by remote enhancers in the cut and Ultrabithorax genes. Genetics 152: 577-593. PubMed citation: 10353901

Rollins, R. A., Korom, M., Aulner, N., Martens, A. and Dorsett, D. (2004). Drosophila Nipped-B protein supports sister chromatid cohesion and opposes the Stromalin/Scc3 cohesion factor to facilitate long-range activation of the cut gene. Mol. Cell. Biol. 24: 3100-3111. PubMed citation: 15060134

Schuldiner, O. et al. (2008), piggyBac-based mosaic screen identifies a postmitotic function for cohesin in regulating developmental axon pruning. Dev. Cell 14: 227-238. PubMed citation: 18267091

Suster, M. L., et al. (2004). Turning behavior in Drosophila larvae: a role for the small scribbler transcript. Genes Brain Behav. 3(5): 273-86. PubMed citation: 15344921

Takagi, S., et al. (1997). Cellular and axonal migrations are misguided along both body axes in the maternal-effect mau-2 mutants of Caenorhabditis elegans. Development 124: 5115-5126. PubMed citation: 9362469

Tonkin, E. T., et al. (2004). NIPBL, encoding a homolog of fungal Scc2-type sister chromatid cohesion proteins and fly Nipped-B, is mutated in Cornelia de Lange syndrome. Nat. Genet. 36: 636-641. PubMed citation: 15146185

Valdeolmillos, A., et al. (2004). Drosophila cohesins DSA1 and Drad21 persist and colocalize along the centromeric heterochromatin during mitosis. Biol. Cell 96: 457-462. PubMed citation: 15325074

Vass, S. et al. (2003). Depletion of Drad21/Scc1 in Drosophila cells leads to instability of the cohesin complex and disruption of mitotic progression. Curr. Biol. 13: 208-218. PubMed citation; Online text

Waizenegger, I. C. et al. (2000. Two distinct pathways remove mammalian cohesin from chromosome arms in prophase and from centromeres in anaphase. Cell 103: 399-410. PubMed citation: 11081627

Warren, W. D., et al. (2000). The Drosophila RAD21 cohesin persists at the centromere region in mitosis. Curr. Biol. 10(22): 1463-6. PubMed citation: 11102811

date revised: 3 February 2008

Home page: The Interactive Fly © 2008 Thomas Brody, Ph.D.