Search

Summary Cohesin organizes DNA into chromatids, regulates enhancer-promoter interactions, and confers sister chromatid cohesion. Its association with chromosomes is regulated by hook-shaped HEAT repeat proteins that bind Scc1, namely Scc3, Pds5, and Scc2. Unlike Pds5, Scc2 is not a stable cohesin constituent but, as shown here, transiently replaces Pds5. Scc1 mutations that compromise its interaction with Scc2 adversely affect cohesin’s ATPase activity and loading. Moreover, Scc2 mutations that alter how the ATPase responds to DNA abolish loading despite cohesin’s initial association with loading sites. Lastly, Scc2 mutations that permit loading in the absence of Scc4 increase Scc2’s association with chromosomal cohesin and reduce that of Pds5. We suggest that cohesin switches between two states: one with Pds5 bound that is unable to hydrolyze ATP efficiently but is capable of release from chromosomes and another in which Scc2 replaces Pds5 and stimulates ATP hydrolysis necessary for loading and translocation from loading sites., Graphical Abstract, Highlights • Among Hawks, Scc2 is sufficient to activate cohesin’s ATPase in the presence of DNA • Scc1 mutants that compromise Scc2 binding reduce ATPase activity and cohesin loading • Scc2 gain-of-function mutations increase its association with chromosomal cohesin • Increased Scc2 association with chromosomal cohesin correlates with a decrease in Pds5, Cohesin switches between two states: one with Pds5 bound that is unable to hydrolyze ATP efficiently but is capable of release from chromosomes and another one in which Scc2 replaces Pds5 and stimulates ATP hydrolysis required for loading and translocation.

Summary Cohesin entraps sister DNAs within tripartite rings created by pairwise interactions between Smc1, Smc3, and Scc1. Because Smc1/3 ATPase heads can also interact with each other, cohesin rings have the potential to form a variety of sub-compartments. Using in vivo cysteine cross-linking, we show that when Smc1 and Smc3 ATPases are engaged in the presence of ATP (E heads), cohesin rings generate a “SMC (S) compartment” between hinge and E heads and a “kleisin (K) compartment” between E heads and their associated kleisin subunit. Upon ATP hydrolysis, cohesin’s heads associate in a different mode, in which their signature motifs and their coiled coils are closely juxtaposed (J heads), creating alternative S and K compartments. We show that K compartments of either E or J type can entrap single DNAs, that acetylation of Smc3 during S phase is associated with J heads, and that sister DNAs are entrapped in J-K compartments., Graphical Abstract, Highlights • Smc1 and Smc3 ATPase heads adopt an engaged (E) and a juxtaposed (J) state in vivo • Smc ATPase heads delimit an Smc (S) and a kleisin (K) compartment • Single DNA molecule can be entrapped inside K compartments of either E or J type • Sister DNAs are entrapped in J-K compartments with J-head Smc3 being acetylated, Sister chromatid cohesion is thought to be conferred by sister DNA entrapment within the cohesin ring. Chapard et al. show that cohesin’s Smc heads interact with each other in two different ways in vivo, allowing sister DNA entrapment between juxtaposed ATPase heads and kleisin of the cohesin complex.

We now take for granted that despite the disproportionate contribution of females to initial growth of their progeny, there is little or no asymmetry in the contribution of males and females to the eventual character of their shared offspring. In fact, this key insight was only established towards the end of the eighteenth century by Joseph Koelreuter's pioneering plant breeding experiments. If males and females supply equal amounts of hereditary material, then the latter must double each time an embryo is conceived. How then does the amount of this mysterious stuff not multiply exponentially from generation to generation? A compensatory mechanism for diluting the hereditary material must exist, one that ensures that if each parent contributes one half, each grandparent contributes a quarter, and each great grandparent merely an eighth. An important piece of the puzzle of how hereditary material is diluted at each generation has been elucidated over the past ten years.

Cohesin's structural maintenance of chromosome 1 (Smc1) and Smc3 are rod-shaped proteins with 50-nm long intra-molecular coiled-coil arms with a heterodimerization domain at one end and an ABC-like nucleotide-binding domain (NBD) at the other. Heterodimerization creates V-shaped molecules with a hinge at their centre. Inter-connection of NBDs by Scc1 creates a tripartite ring within which, it is proposed, sister DNAs are entrapped. To investigate whether cohesin's hinge functions as a possible DNA entry gate, we solved the crystal structure of the hinge from Mus musculus, which like its bacterial counterpart is characterized by a pseudo symmetric heterodimeric torus containing a small channel that is positively charged. Mutations in yeast Smc1 and Smc3 that together neutralize the channel's charge have little effect on dimerization or association with chromosomes, but are nevertheless lethal. Our finding that neutralization reduces acetylation of Smc3, which normally occurs during replication and is essential for cohesion, suggests that the positively charged channel is involved in a major conformational change during S phase. © 2011 European Molecular Biology Organization | All Rights Reserved.

Sequencing DNA fragments associated with proteins following in vivo cross-linking with formaldehyde (known as ChIP-seq) has been used extensively to describe the distribution of proteins across genomes. It is not widely appreciated that this method merely estimates a protein's distribution and cannot reveal changes in occupancy between samples. To do this, we tagged with the same epitope orthologous proteins in Saccharomyces cerevisiae and Candida glabrata, whose sequences have diverged to a degree that most DNA fragments longer than 50 bp are unique to just one species. By mixing defined numbers of C. glabrata cells (the calibration genome) with S. cerevisiae samples (the experimental genomes) prior to chromatin fragmentation and immunoprecipitation, it is possible to derive a quantitative measure of occupancy (the occupancy ratio - OR) that enables a comparison of occupancies not only within but also between genomes. We demonstrate for the first time that this 'internal standard' calibration method satisfies the sine qua non for quantifying ChIP-seq profiles, namely linearity over a wide range. Crucially, by employing functional tagged proteins, our calibration process describes a method that distinguishes genuine association within ChIP-seq profiles from background noise. Our method is applicable to any protein, not merely highly conserved ones, and obviates the need for the time consuming, expensive, and technically demanding quantification of ChIP using qPCR, which can only be performed on individual loci. As we demonstrate for the first time in this paper, calibrated ChIP-seq represents a major step towards documenting the quantitative distributions of proteins along chromosomes in different cell states, which we term biological chromodynamics.

In mammalian females, germ cells remain arrested as primordial follicles. Resumption of meiosis is heralded by germinal vesicle breakdown, condensation of chromosomes, and their eventual alignment on metaphase plates. At the first meiotic division, anaphase-promoting complex/cyclosome associated with Cdc20 (APC/C) activates separase and thereby destroys cohesion along chromosome arms. Because cohesion around centromeres is protected by shugoshin-2, sister chromatids remain attached through centromeric/pericentromeric cohesin. We show here that, by promoting proteolysis of cyclins and Cdc25B at the germinal vesicle (GV) stage, APC/C associated with the Cdh1 protein (APC/C) delays the increase in Cdk1 activity, leading to germinal vesicle breakdown (GVBD). More surprisingly, by moderating the rate at which Cdk1 is activated following GVBD, APC/C creates conditions necessary for the removal of shugoshin-2 from chromosome arms by the Aurora B/C kinase, an event crucial for the efficient resolution of chiasmata., A.R., R.B.M., and M. Hopkins were supported by PhD fellowships from the Boehringer Ingelheim Fonds, Barbour Foundation, and EPSRC (EP/G03706X/1), respectively. A.M.P. is supported by Ministerio de Economia y Competitividad (MINECO) (grant number: BFU-2014-59307); M. Herbert is funded by the Medical Research Council (MR/J003603/1), Wellcome Trust (096919), and European Community’s Horizon 2020 Research and Innovation Programme under grant agreement 634113 (GermAge); and B.N. is supported by a BBSRC Strategic LoLa grant (BB/M00354X/1). The European Community’s Seventh Framework MitoSys (241548), Medical Research Council (84673), and Wellcome Trust (019859/Z/10/Z) funded this project. Open Access funded by Wellcome Trust.

Sister chromatid cohesion (cohesion) is a process that ensures accurate segregation of genetic material into daughter cells during mitosis in eukaryotic cells. Cohesion is mediated by a highly conserved ring-shaped protein complex called cohesin. Cohesion is thought to arise from the topological entrapment of replicated sister DNAs inside the cohesin ring. Although cohesin can load onto, and topologically entrap individual DNAs prior to DNA replication, cohesion is established during S-phase in concert with DNA replication. Analysis of small circular DNAs in yeast has revealed that cohesion establishment can occur by two pathways operating in parallel. Firstly, cohesin can load onto replicated sister DNAs de novo in a manner that is dependent on the cohesin loading protein Scc2. Secondly, there exists a ‘conversion’ pathway in which cohesin binds to unreplicated DNA during G1, and remains associated with DNA throughout S-phase during which it is converted into cohesive cohesin. This process is independent of Scc2 but requires a group of replisome-associated factors which include Ctf4. Whether this cohesin conversion pathway generates cohesion of yeast chromosomes, and the molecular mechanisms underlying conversion, remain largely unknown. It is thought that opening of cohesin’s Smc1-Smc3 hinge is crucial for entrapment of individual DNAs, but whether this is the case for entrapment of sister DNAs is currently unclear. I have developed a unique assay to measure cohesion generated solely by cohesin rings associated with DNA from G1. Using this assay, we show for the first time that conversion of non-cohesive cohesin to cohesive cohesin occurs on S. cerevisiae chromosomes during DNA replication, and that this process does not require opening of the Smc1-Smc3 hinge. Furthermore, the conversion observed was significantly reduced by deletion of Ctf4 but not significantly impacted by inactivation of the cohesin loader Scc2. These data provide crucial mechanistic insight into the process of cohesin conversion, and unequivocally demonstrate that while hinge opening is required for loading of cohesin rings onto DNA, it is not required for conversion of these rings into cohesive forms.

Cohesin is essential for the spatial organization of the genome through its ability to extrude loops on the DNA. Structural information indicates that the cohesin’s Smcs fold at the elbow, bringing the hinge domain close to the ATPase heads and also can dissociate from each other to adopt an unzipped coils conformation. In all proposed loop extrusion models, cohesin needs to undergo changes to its conformational state, with unzipping or unfolding of the complex being necessary for its ATP hydrolysis driven loop extrusion. In this work, I investigate the role of cohesin’s conformation in relation to its ATPase activity. Specifically I am looking at its the folding and the zipping up of the coiled coils. I examine the changes that occur in those conformational states during cohesin’s ATP hydrolysis cycle while also considering their regulatory role. My research shows that both unzipping and unfolding occur during the hydrolysis cycle, while only unzipping is necessary for ATP hydrolysis. I find that the zipped-up coils inhibition is lost in cohesin trimers with mutations at the hinge that do not fold. The unfolded cohesin trimers hydrolyse ATP more readily than WT, indicating that the folded state is required for the stability of the zipping up of the coils. My results lead me to the conclusion that cohesin’s rigid APO state inhibits ATP hydrolysis due to the zipped up coils that are stabilized by the folding at the elbow. ATP hydrolysis therefore, requires Scc2 and DNA to ensure the unzipping of the coils in the clamped state.

Cohesin is an SMC complex essential for genome organisation and the accurate segregation of sister chromatids in mitosis. The activity of this ATP dependent ring shape trimer composed of two SMC proteins and a kleisin subunit relies on two additional complexes. While the loading complex made of Scc2 and Scc4 is essential for the recruitment of cohesin at the loading sites, the releasing complex comprised of Pds5, Scc3 and Wapl is necessary to adjust the turnover of cohesin on DNA. During replication, Smc3 is acetylated and cohesion is established between the two newly replicated DNA molecules. Scc2 is essential for the ATPase activity of cohesin, its loading on chromatin and the genome reorganisation by DNA loop extrusion. In order to better understand the biological role of Scc2, the interaction between Scc2 and the different subunits of the cohesin complex was investigated using genetic mutations combined with in vivo and in vitro protein crosslinking. Scc2 makes extensive contacts with at least three regions of Scc1 whose deletion cause defects in loading and translocation of cohesin. Scc2 competes with Pds5 in order to stimulate the ATPase activity of the complex, but the mutation of the binding site in Scc1, common to Pds5 and Scc2, is not sufficient to abolish cohesin loading. By using different crosslinking strategies, accurate binding sites could be identified between Scc2 and Scc1. Scc2 was also found to interact with a conserved loop in the head of Smc1 that is essential for its engagement with Smc3’s head and ATP hydrolysis. Finally, crosslinking experiments confirmed the existence in vivo of the folded configuration of cohesin that brings the hinge into contact with the coiled coils. This configuration was observed throughout the cell cycle, independently of cohesin acetylation, suggesting that Scc2 is not involved in its stabilisation.

Cohesin is a ring like SMC-kleisin complex (S-K ring) that mediates DNA interactions between chromosomes. It is responsible for the phenomenon of sister chromatid cohesion (SCC) and ensures the faithful segregation of chromosomes during mitosis. Cohesin also mediates interactions within chromosomes by organising DNA into loops in both interphase and mitosis. SCC is thought to be mediated by co-entrapment of sister DNAs within a single cohesin S-K ring. Cohesin will also entrap individual chromosomes within its S-K ring during G1, and it is thought this activity is a pre-requisite to cohesion establishment during S phase. To improve our understanding of the mechanism of the entrapment of DNA by S-K rings, I developed a protocol to measure this activity in vitro. By using a cysteine-cysteine crosslinking protocol, I was able to covalently circularise the cohesin ring on DNA and through this create protein-DNA catenations that can resist protein denaturation. This revealed that S-K entrapment was dependent on Scc3 and ATP binding, and was stimulated by Scc2 and ATP hydrolysis. I found both Scc2 and Scc3 are DNA binding proteins and discover that this DNA binding activity is essential for cohesin’s ability to entrap DNA within its S-K ring. I then set out to identify where within the cohesin ring DNA was entrapped, and through this managed to identify a novel mode of DNA interaction, whereby DNA becomes entrapped within two compartments created through cohesin binding to ATP. Entrapment within these two compartments was not associated with passage of DNA into S-K rings, and I propose entrapment of this nature may represent the first step in cohesin’s topological association with DNA.

Missegregation of chromosomes during meiosis leads to formation of aneuploid eggs. Estimates suggest that in humans, about 10-30% of fertilised eggs and one-third of all miscarriages are aneuploid. Accurate chromosome segregation depends on the coordination between stepwise cohesion resolution and attachments of homologous chromosomes through kinetochores to microtubules, emanating from opposite poles of the cell. The Spindle Assembly Checkpoint (SAC) monitors microtubule-kinetochore attachments and prevents resolution of cohesin complexes by inhibiting the ubiquitin ligase APC/CCdc20 until all aberrant microtubule-kinetochore attachments have been rectified by an Aurora Kinase-dependent error correction machinery. During meiosis, these pathways work in seamless coordination to achieve balanced segregation of the genome at the first meiotic division.The cross-talk between different cell cycle pathways requires members with shared affiliations. During my DPhil studies, I worked on understanding the role of two such proteins, namely Bub1 (budding uninhibited by benzimidazoles 1) and Sgol2 (Shugoshin-like protein 2) in mouse oocytes.During the first meiotic division, Bub1 maintains the SAC, and through its kinase activity, Bub1 recruits Sgol2 to kinetochores to protect centromeric cohesion. This recruitment is essential for two rounds of chromosomes segregation in meiosis. Thus, Bub1 localisation at kinetochores can coordinate the timing of anaphase with the centromeric cohesion protection.During the first meiotic division, Sgol2 protects centromeric cohesion by recruiting PP2A to kinetochores, accelerates cohesin resolution by silencing the SAC through its interaction with PP2A and Mad2, and also promotes biorientation and congression of homologous chromosomes by its interaction with MCAK and through dephosphorylation of Aurora B/C kinase substrates at kinetochores.This research revealed that Bub1 and Sgol2 can regulate anaphase by linking multiple cell cycle pathways that work together to achieve faithful chromosomes segregation in mammalian meiosis.

Cohesion between sister chromatids that are newly formed in S-phase is crucial to ensure the fidelity of chromosome segregation. The cohesin complex mediates cohesion and is comprised of two SMC proteins, Smc1 and Smc3, with a kleisin subunit, Scc1. The SMCs interact via their hinge domains to form a v-shaped heterodimer, with Scc1 forming a bridge between their head domains to produce a ring structure that topologically entraps sister chromatids. In order to confer cohesion, cohesin needs to be loaded onto DNA, a process that is poorly understood. It is known that cohesin loading depends on a loading complex comprised of Scc2 and Scc4, ATP hydrolysis and presumably opening of at least one of the interfaces of the cohesin ring. Using the recently developed technique, quantitative ChIP-seq, we show that Pds5, a cohesin-associated protein, is displaced at centromeres from the cohesin ring during the loading reaction. This event is accompanied by the engagement of SMC ATPase heads in the presence of ATP. Upon ATP hydrolysis, the loading reaction completes with DNA entrapment, translocation to the pericentromeric sequences and the displacement of Scc2 by Pds5. We also describe mutations in Scc2, Smc1 and histone proteins H2A and H2B that enable loading in the absence of Scc4. The mutations in Scc2 alter the dynamics of the association between the loading complex and cohesin, reducing dissociation during the translocation step. All of the mutations are able to restore loading along chromosome arms but not at centromeres where Scc4 has a particular role. However, the Smc1 mutation hinders loading at the centromere, even in the presence of Scc4, due to a defect in the second step of the loading reaction, namely translocation driven by ATP hydrolysis. We suggest that this discrepancy in functionality at the arms and centromere is due to changes in the hinge accompanying both the first and second steps. The rate limiting reaction is the first step at arms and the second at the centromeres. The histone mutants show that nucleosome occupancy plays an important role, at least on chromosome arms, but we find no evidence that this involves the RSC complex.

The accurate inheritance of genetic material depends upon the establishment and maintenance of sister chromatid cohesion. Replicated chromosomes are topologically encircled by the large, tripartite protein complex cohesin, allowing bi-orientation in mitosis. To entrap and reversibly dissociate from DNA, the annular complex structure must be disrupted at either the hinge domain between Smc1 and Smc3, or the interfaces created by the kleisin subunit Scc1 bridging the two ABC-like ATPase domains. The aim of this work was to characterise the cohesin complex loading and releasing mechanisms by examining the biochemical requirements for these processes. Although the identity of a chromosomal cofactor could not be assigned, the loading reaction was found to necessitate engagement of ATPase domains in an ATP-dependent manner. Notions of allosteric modulation of ATP binding and NBD engagement by acetylation were discredited. Likewise, a direct and stable physical association of hinge domains with NBDs was shown to be an unlikely conformational intermediate in a reaction thought to promote hinge opening for loading of cohesin onto DNA. The Smc3-Scc1 kleisin interface might be exploited during the opposing process of release of cohesin from DNA. Therefore, a novel protein- protein cross-linking system was adapted for use in S. cerevisiae, with a view to (1) confirming the well-founded role of hinge dissociation in topological entrapment, and (2) validating the Smc3- Scc1 interface as the recently conjectured exit gate. Despite promising preliminary kinetics in vitro, the SpyTag-SpyCatcher system was considerably less efficient in vivo. It was thus deemed unsuitable in its current format for investigating the process of interface dissociation in live cells. Finally, the large, cohesin complex-associated HEAT repeat protein Pds5 has been either speculated or shown to participate in all of the aforementioned processes, potentially modulating the hinge and Smc-kleisin interfaces. Acting as a regulatory node in cohesin function, it was pursued as an informative target for structural studies. Although no diffractive material could be obtained, Pds5 was confirmed to bind a short, N-terminal sequence of Scc1 and was in turn bound at its N- terminus by Wapl. Together, these findings contribute to defining the structures and states of the cohesin complex.

Sister chromatid cohesion is essential for accurate chromosome segregation. Cohesion is generated by cohesin, a conserved multi-subunit protein complex composed of four core subunits: Smc1, Smc3, Scc1, and Scc3. Cohesin holds sister chromatids together in mitotic cells starting from S-phase, when DNA replicates, until their separation at the onset of anaphase where its Scc1 subunit is cleaved. In budding yeast, most Scc1 is destroyed by cleavage at anaphase and is only re-synthesised in late G1, whereupon it associates with the unreplicated chromatin. Although sister chromatid cohesion is known to be mediated by a topological interaction of cohesin complexes around sister DNAs, the nature of cohesin`s interaction with chromatin before DNA replication remains to be elucidated. My project aims to develop a new system in order to find out whether ‘non-cohesive’ cohesin interacts with chromatin topologically. This is important for two main reasons. Firstly, understanding the physical nature of cohesin’s interaction with chromatin before DNA replication is essential for determining how cohesion is established during DNA replication. Another reason is that most cohesin in multicellular organisms is associated with the unreplicated chromatin of post mitotic cells where it regulates transcription. How cohesin mediates gene expression is unknown. Understanding how cohesin binds unreplicated chromatin may therefore bring insights into the mechanisms by which cohesin complex performs its non-canonical functions. In order to address this, we needed a situation where cohesin is already loaded onto chromosomes, but either DNA replication or cohesion establishment is prevented. Therefore, we used a temperature sensitive allele of Eco1 (required for establishment of cohesion). Quantitative measurement of cohesin levels on chromosomes in either wild type allele or temperature sensitive allele of Eco1 showed that the amount of cohesin associated with centromeric and inner pericentromeric regions in both strains are almost indistinguishable from each other throughout the whole cell cycle. Despite normal levels of cohesin, we confirmed by minichromosomal assay that no sister chromatid cohesion is established in the absence of functional Eco1 protein. If “non-cohesive” cohesin interacts with the chromatin in a topological manner when there is no sister chromatid cohesion, then its association with chromatin should be resistant to denaturing conditions in the presence of a modified version of the cohesin complex that can be covalently circularized. To test this prediction, a cross-linkable cohesin molecule was needed, which should be resistant to SDS denaturation and should not have major cohesion defects due to the modifications making it to be cross-linkable. The previously created cross-linkable cohesin molecule had cohesion defects due to the presence of Smc3-Scc1 fusion protein. In addition, this fusion alone could bypass the requirement for Eco1, and therefore we could not test how “non-cohesive” cohesin interacts with chromatin, using this version of cross-linkable cohesin complex. We tried two different methods to conditionally close Smc3/Scc1 interface in a way resistant to protein-denaturants and create a new cross-linkable cohesin complex. In our first attempt, the C-terminus of Smc3 and the N-terminus of Scc1 were fused to FRB and FKBP12 respectively, proteins that can form a complex upon addition of rapamycin. Crystal structure of the ternary complex of FKP12/rapamycin/FRB enabled us to design cysteine pairs for the crosslinking of FRB and FKBP12 only in the presence of rapamycin. A more efficient in vivo crosslinking was achieved between the Smc3 and Scc1 in our second attempt. Amino acids within the coiled coil region of Smc3 were replaced by the unnatural photo-cross-linkable amino acid ρ-benzoyl-phenylalanine that can be induced to form covalent bonds with neighbouring proteins (T.Gligoris, unpublished data). Photo and chemically cross-linkable interfaces of cohesin were then integrated with each other to generate a new version of cross-linkable cohesin molecule.

Cohesin, a ring-shaped complex composed of four subunits, is an essential player involved in timely and accurate segregation of genetic material at mitosis and meiosis. Cohesin performs a highly conserved role by topologically embracing sister chromatids until their concerted disjunction at anaphase, when the α-kleisin subunit of the ring is irreversibly cleaved by separase. A central part of the cohesin cycle is its cleavage-independent removal from chromatin through the action of the releasing complex. In more complex eukaryotes, stabilisation of cohesin on DNA depends on the presence of sororin, a small functionally conserved protein whose association with cohesin is thought to counteract the releasing reaction. The first part of this thesis aimed to address the discrepancy of a sororin orthologue never having been identified in S. cerevisiae. This was undertaken using an imaging-based approach to screen a subset of the yeast genome for proteins with cohesin-like localisation. The screen was streamlined such that candidate genes had periodically cycling mRNA transcripts, similarly to cohesin subunits. No novel cohesin-associated proteins could be identified, although this could be in part due to technical limitations of the screen. The precise molecular details of cohesin’s function and regulation remain poorly understood. A conclusive way to address many questions about cohesion as a whole would be to set up an in vitro assay capable of reconstructing cohesin loading, cohesion establishment, and the releasing reaction. The second part of this thesis aimed to establish a reliable purification protocol for the S. cerevisiae cohesin complex using a bacterial expression system for use in such biochemical assays. While purification of all components of the cohesin trimer (Smc1, Smc3, Scc1) was achieved successfully, the complex could not be confirmed to assemble functionally, as determined by chemical crosslinking of the Smc3/Scc1 interface.

Cohesin is a highly conserved multisubunit complex that holds sister chromatids together in mitotic cells. 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, we have created fruit flies 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. © 2008 Elsevier Inc. All rights reserved.