The many functions of cohesin—different rings to rule them all?

EMBO J 31 9, 2076–2089 March132012EMBO J 31 9, 2090–2102 March132012It is well known that somatic and germ cells use different cohesin complexes to mediate sister chromatid cohesion, but why different isoforms of cohesin also co-exist within somatic vertebrate cells has remained a mystery. Two papers in this issue of The EMBO Journal have begun to address this question by analysing mouse cells lacking SA1, an isoform of a specific cohesin subunit.When one cell divides into two, many things have to go right for the two daughter cells to receive identical copies of their mother cell''s genome. It has long been recognized that sister chromatid cohesion, the physical connection established during DNA replication between newly synthesized sister DNA molecules, is one of these essential prerequisites for proper chromosome segregation. It is this cohesion that enables the bi-orientation of chromosomes on the mitotic or meiotic spindle, and thus makes their symmetrical segregation possible. Cohesion is mediated by cohesin, a multi-subunit protein complex, which is thought to connect sister DNA molecules by embracing them as a ring (Figure 1; reviewed in Peters et al, 2008). It is well established that cohesin complexes differ between somatic and germ cells, where they are needed for the proper separation of sister chromatids and of homologous chromosomes, respectively. What has been largely ignored, however, is that even within somatic vertebrate cells there are different forms of cohesin, containing mutually exclusive variable subunits: either SA1 or the closely related SA2 protein (also known as STAG1 and STAG2, respectively), and either Pds5A or the related Pds5B subunit (Peters et al, 2008). Why is that? Two papers from the Losada lab (Remeseiro et al, 2012a, 2012b) have begun to address this question by generating mouse cells lacking the SA1 gene, revealing unexpected insights into the functions of SA1 subunit-containing cohesin complexes (cohesin-SA1).Open in a separate windowFigure 1Schematic drawing illustrating how the SA1 and SA2 proteins interact in a mutually exclusive manner with three core subunits of cohesin (Smc1, Smc3, Rad21) that form a ring-like structure. It has been proposed that these complexes mediate cohesion by trapping the two sister DNA molecules inside the cohesin ring (above), and that cohesin rings might affect chromatin structure by forming or stabilizing intra-chromatid loops (below). Cohesin is thought to influence gene regulation at least in part by mediating chromatin looping.Although cohesin is best known for its role in sister chromatid cohesion, it is clearly also needed for homologous recombination-mediated DNA repair and for gene regulation. Much of what we know about these functions comes from studies in yeast and fruit flies, organisms with only a single SA1/SA2-related mitotic subunit (Scc3 in budding yeast), and only a single Pds5 subunit. It is therefore plausible that, like many other genes during vertebrate evolution, SA1/SA2 and Pds5A/Pds5B have arisen by gene duplication to constitute paralogs, with functional differences between them assumed to be subtle. Consistently, absence of either Pds5A or Pds5B causes only mild, if any, defects in sister chromatid cohesion, and mice lacking either protein can develop to term, although they die shortly after birth owing to multiple organ defects (Zhang et al, 2007, 2009). First indications that the situation may be different for the Scc3-related subunits came from Canudas and Smith (2009), who reported that RNAi depletion of SA1 and SA2 from HeLa cells caused defects in telomere and centromere cohesion, respectively. The generation of mice lacking either one or both alleles of the SA1 gene has now allowed a more systematic and thorough analysis of SA1 function (Remeseiro et al, 2012a, 2012b).One of the most striking results obtained in these studies is that most mice lacking SA1 die around day 12 of embryonic development, clearly showing that the function of SA1 cannot be fulfilled by SA2, despite the fact that SA2 is substantially more abundant in somatic cells than SA1 (Holzmann et al, 2010). What could this SA1-specific function be? Losada and colleagues report observations, which imply that SA1 does not have just one, but possibly several important functions in different processes. First, the authors confirm the previous observation that SA1 is required for cohesion specifically at telomeres, while likely collaborating with SA2 in chromosome arms or centromeric regions. Furthermore, telomeres have an unusual morphology in mitotic chromosomes lacking SA1 (Remeseiro et al, 2012a), reminiscent of a fragile-site phenotype previously reported in telomeres with DNA replication defects (Sfeir et al, 2009), and SA1 is indeed required for efficient telomere duplication. Depletion of sororin, a protein that is required for cohesin''s ability to mediate sister chromatid cohesion, also causes a fragile-site phenotype at telomeres. These findings imply that SA1''s role in telomere cohesion is important for efficient telomere replication, perhaps, as the authors speculate, because telomere cohesion may help to stabilize or re-start stalled replication forks, or because cohesion-dependent homologous recombination might be involved in repair of DNA double strand breaks created by collapsed replication forks. Interestingly, cells lacking SA1 frequently show chromosome bridges in anaphase, often fail to divide, and either die or become bi-nucleated. The exact origin of chromosome bridges is difficult to determine, but previous studies have found such bridges often associated with fragile sites on chromosomes; treatment with low doses of DNA replication inhibitors was shown to increase the frequency of such bridges (Chan et al, 2009), and similar observations were indeed made by Remeseiro et al (2012a) in mouse embryonic fibroblasts. It is therefore plausible that the telomere cohesion defect observed in SA1-lacking cells leads to incomplete telomere replication, which in turn results in the formation of anaphase chromosome bridges and subsequent cytokinesis defects. Losada and colleagues further speculate that these chromosome segregation defects could underlie the increased frequency of spontaneous development of various tumours in mice containing just one instead of two SA1 alleles (Remeseiro et al, 2012a). This is an attractive interpretation since tetraploidy and aneuploidy are thought to contribute to the rate with which tumour cells can evolve; however, Losada and colleagues report SA1 deficiency to cause defects also in other cohesin functions, which may therefore as well contribute to tumour formation.To further understand why SA1 cannot be fulfilled by SA2, Losada and colleagues also analysed the distribution of these proteins in the non-repetitive parts of the mouse genome by chromatin immunoprecipitation coupled to deep sequencing (ChIP-seq). The results of these experiments, published in the second of the two papers (Remeseiro et al, 2012b), raise the interesting possibility that cohesin-SA1 associates more frequently with gene promoters than cohesin-SA2. However, the fact that different antibodies have to be used for any ChIP-based comparison of the distribution of two proteins makes it difficult to know to what degree observed differences might be due to different antibody efficiency. Obviously, such limitations do not exist if the distribution of one and the same protein is analysed under different conditions, and in such an experimental setting, Remeseiro et al indeed make some striking observations. When SA1 is absent, SA2 does not detectably change in abundance, but its distribution in the genome does, in that more than half of all SA2-binding sites in SA1-deficient cells differ from those bound in wild-type cells. Most SA2-binding sites in SA1-deficient cells are in intergenic regions, and CTCF, a zinc finger protein often co-localizing with cohesin and implicated in its gene regulation function (Peters et al, 2008), appears to be absent at many of these sites. It presently remains a mystery why cohesin-SA2 changes its distribution so dramatically in the absence of SA1, but the observation that gene promoters are more frequently occupied by cohesin in the presence of SA1 than in its absence raises the possibility that cohesin-SA1 may have a specific role in gene regulation. This possibility is particularly interesting in light of a recent study that found hardly any change in gene expression upon re-expression of SA2 in SA2-deficient human glioblastoma cells (Solomon et al, 2011), despite the fact that cohesin is thought to regulate numerous genes. With this in mind, Remeseiro et al analysed gene expression in mouse cells and indeed found 549 genes to be mis-regulated in the absence of SA1, in striking contrast to the above-mentioned comparison of human SA2-deficient or proficient cells that found only 19 genes to change in expression levels (Solomon et al, 2011). Obviously direct comparisons will be essential to analyse further the specific roles of SA1 and SA2 in gene regulation, but the current evidence raises the interesting possibility that SA1 may have a particularly important role in gene regulation, whereas cohesin-SA2 is dedicated to creating arm and centromeric cohesive structures for chromosome segregation.That is not to say that cohesin-SA1 cannot mediate sister chromatid cohesion. It almost certainly can, as it is essential for cohesion at telomeres (Canudas and Smith, 2009; Remeseiro et al, 2012a). Likewise, it would be wrong to assume that we now fully understand why SA1 and SA2 co-exist in somatic vertebrate cells, and what their precise functions is. There are many things we do not understand. For example, if SA2 has little or no role in gene regulation, as the Solomon et al (2011) study implies, why does SA2 nevertheless interact directly with CTCF (Xiao et al, 2011), its gene regulation collaborator? How do cohesin-SA1 and cohesin-SA2 complexes further differ in their genomic distributions and their functions depending on whether they contain either Pds5A or Pds5B, constituting really not just two but four distinct cohesin complexes? The work by Losada and colleagues represents an important step towards understanding these questions, but there is still a long and presumably exciting way to go to understand how different cohesin complexes control the mammalian genome.     

Protection by natural IgG: a sweet partnership with soluble lectins does the trick!

EMBO J 32: 2905–2919 10.1038/emboj.2013.199; published online September032013Some B cells of the adaptive immune system secrete polyreactive immunoglobulin G (IgG) in the absence of immunization or infection. Owing to its limited affinity and specificity, this natural IgG is thought to play a modest protective role. In this issue, a report reveals that natural IgG binds to microbes following their opsonization by ficolin and mannan-binding lectin (MBL), two carbohydrate receptors of the innate immune system. The interaction of natural IgG with ficolins and MBL protects against pathogenic bacteria via a complement-independent mechanism that involves IgG receptor FcγRI expressing macrophages. Thus, natural IgG enhances immunity by adopting a defensive strategy that crossovers the conventional boundaries between innate and adaptive microbial recognition systems.The adaptive immune system generates protective somatically recombined antibodies through a T cell-dependent (TD) pathway that involves follicular B cells. After recognizing antigen through the B-cell receptor (BCR), follicular B cells establish a cognate interaction with CD4⁺ T follicular helper (T_FH) cells and thereafter either rapidly differentiate into short-lived IgM-secreting plasmablasts or enter the germinal centre (GC) of lymphoid follicles to complete class switch recombination (CSR) and somatic hypermutation (SHM) (Victora and Nussenzweig, 2012). CSR from IgM to IgG, IgA and IgE generates antibodies with novel effector functions, whereas SHM provides the structural correlate for the induction of affinity maturation (Victora and Nussenzweig, 2012). Eventually, this canonical TD pathway generates long-lived bone marrow plasma cells and circulating memory B cells that produce protective class-switched antibodies capable to recognize specific antigens with high affinity (Victora and Nussenzweig, 2012).In addition to post-immune monoreactive antibodies, B cells produce pre-immune polyreactive antibodies in the absence of conventional antigenic stimulation (Ehrenstein and Notley, 2010). These natural antibodies form a vast and stable repertoire that recognizes both non-protein and protein antigens with low affinity (Ehrenstein and Notley, 2010). Natural antibodies usually emerge from a T cell-independent (TI) pathway that involves innate-like B-1 and marginal zone (MZ) B cells. These are extrafollicular B-cell subsets that rapidly differentiate into short-lived antibody-secreting plasmablasts after detecting highly conserved microbial and autologus antigens through polyreactive BCRs and nonspecific germline-encoded pattern recognition receptors (Pone et al, 2012; Cerutti et al, 2013).The most studied natural antibody is IgM, a pentameric complement-activating molecule with high avidity but low affinity for antigen (Ehrenstein and Notley, 2010). In addition to promoting the initial clearance of intruding microbes, natural IgM regulates tissue homeostasis, immunological tolerance and tumour surveillance (Ochsenbein et al, 1999; Zhou et al, 2007; Ehrenstein and Notley, 2010). Besides secreting IgM, B-1 and MZ B cells produce IgG and IgA after receiving CSR-inducing signals from dendritic cells (DCs), macrophages and neutrophils of the innate immune system (Cohen and Norins, 1966; Cerutti et al, 2013). In humans, certain natural IgG and IgA are moderately mutated and show some specificity, which may reflect the ability of human MZ B cells to undergo SHM (Cerutti et al, 2013). Yet, natural IgG and IgA are generally perceived as functionally quiescent.In this issue, Panda et al show that natural IgG bound to a broad spectrum of bacteria with high affinity by cooperating with ficolin and MBL (Panda et al, 2013), two ancestral soluble lectins of the innate immune system (Holmskov et al, 2003). This binding involved some degree of specificity, because it required the presence of ficolin or MBL on the microbial surface as well as lower pH and decreased calcium concentration in the extracellular environment as a result of infection or inflammation (see Figure 1).Open in a separate windowFigure 1Ficolins and MBL are produced by hepatocytes and various cells of the innate immune system and opsonize bacteria after recognizing conserved carbohydrates. Low pH and calcium concentrations present under infection-inflammation conditions promote the interaction of ficolin or MBL with natural IgG on the surface of bacteria. The resulting immunocomplex is efficiently phagocytosed by macrophages through FcγR1 independently of the complement protein C3, leading to the clearance of bacteria.Ficolins and MBL are soluble pattern recognition receptors that opsonize microbes after binding to glycoconjugates through distinct carbohydrate recognition domain (CRD) structures (Holmskov et al, 2003). While ficolins use a fibrinogen domain, MBL and other members of the collectin family use a C-type lectin domain attached to a collagen-like region (Holmskov et al, 2003). Similar to pentraxins, ficolins and MBL are released by innate effector cells and hepatocytes, and thus may have served as ancestral antibody-like molecules prior to the inception of the adaptive immune system (Holmskov et al, 2003; Bottazzi et al, 2010). Of note, MBL and the MBL-like complement protein C1q are recruited by natural IgM to mediate complement-dependent clearance of autologous apoptotic cells and microbes (Holmskov et al, 2003; Ehrenstein and Notley, 2010). Panda et al found that a similar lectin-dependent co-optation strategy enhances the protective properties of natural IgG (Panda et al, 2013).By using bacteria and the bacterial glycan N-acetylglicosamine, Panda et al show that natural IgG isolated from human serum or T cell-deficient mice interacted with the fibrinogen domain of microbe-associated ficolins (Panda et al, 2013). The resulting immunocomplex was phagocytosed by macrophages via the IgG receptor FcγRI in a complement-independent manner (Panda et al, 2013). The additional involvement of MBL was demonstrated by experiments showing that natural IgG retained some bacteria-binding activity in the absence of ficolins (Panda et al, 2013).Surface plasmon resonance provided some clues regarding the molecular requirements of the ficolin–IgG interaction (Panda et al, 2013), but the conformational changes required by ficolin to interact with natural IgG remain to be addressed. In particular, it is unclear what segment of the effector Fc domain of natural IgG binds to ficolins and whether Fc-associated glycans are involved in this binding. Specific glycans have been recently shown to mitigate the inflammatory properties of IgG emerging from TI responses (Hess et al, 2013) and this process could implicate ficolins and MBL. Moreover, it would be important to elucidate whether and how the antigen-binding Fab portion of natural IgG regulates its interaction with ficolins and MBL.The in vivo protective role of natural IgG was elegantly demonstrated by showing that reconstitution of IgG-deficient mice lacking the CSR-enzyme activation-induced cytidine deaminase with natural IgG from T cell-insufficient animals enhanced resistance to pathogenic Pseudomonas aeruginosa (Panda et al, 2013). This protective effect was associated with reduced production of proinflammatory cytokines, occurred independently of the complement protein C3 and was impaired by peptides capable to inhibit the binding of natural IgG to ficolin (Panda et al, 2013). Additional in vivo studies will be needed to determine whether natural IgG exerts protective activity in mice lacking ficolin, MBL or FcγRI, and to ascertain whether these molecules also enhance the protective properties of canonical or natural IgG and IgA released by bone marrow plasma cells and mucosal plasma cells, respectively.In conclusion, the findings by Panda et al show that natural IgG adopts ‘crossover'' defensive strategies that blur the conventional boundaries between the innate and adaptive immune systems. The sophisticated integration of somatically recombined and germline-encoded antigen recognition systems described in this new study shall stimulate immunologists to further explore the often underestimated protective virtues of our vast natural antibody repertoire. This effort may lead to the development of novel therapies against infections.     

Arabidopsis Separase Functions beyond the Removal of Sister Chromatid Cohesion during Meiosis

Separase is a capase family protease that is required for the release of sister chromatid cohesion during meiosis and mitosis. Proteolytic cleavage of the α-kleisin subunit of the cohesin complex at the metaphase-to-anaphase transition is essential for the proper segregation of chromosomes. In addition to its highly conserved role in cleaving the α-kleisin subunit, separase appears to have acquired additional diverse activities in some organisms, including involvement in mitotic and meiotic anaphase spindle assembly and elongation, interphase spindle pole body positioning, and epithelial cell reorganization. Results from the characterization of Arabidopsis (Arabidopsis thaliana) separase (ESP) demonstrated that meiotic expression of ESP RNA interference blocked the proper removal of cohesin from chromosomes and resulted in the presence of a mixture of fragmented chromosomes and intact bivalents. The presence of large numbers of intact bivalents raised the possibility that separase may also have multiple roles in Arabidopsis. In this report, we show that meiotic expression of ESP RNA interference blocks the removal of cohesin during both meiosis I and II, results in alterations in nonhomologous centromere association, disrupts the radial microtubule system after telophase II, and affects the proper establishment of nuclear cytoplasmic domains, resulting in the formation of multinucleate microspores.The proper segregation of chromosomes during mitosis and meiosis is dependent on the systematic formation and subsequent removal of sister chromatid cohesion, which is required for homologous chromosome pairing, recombination, and repair (for review, see Onn et al., 2008; Peters et al., 2008). It is also required for the pairwise alignment of chromosomes on the metaphase I spindle and for the generation of tension across centromeres, thereby ensuring their bipolar attachment. In mitosis, cohesion is maintained by the cohesin complex, which consists of four evolutionally conserved proteins: Sister Chromatid Cohesion1 (SCC1), SCC3, Structural Maintenance of Chromosome1 (SMC1), and SMC3 (for review, see Nasmyth and Haering, 2005). During meiosis, SCC1 is largely replaced by its meiotic homolog REC8.The establishment of sister chromatid cohesion in yeast involves a multistep process (Milutinovich et al., 2007) that begins during telophase of the previous cell cycle when cohesin subunits associate with the chromatin, ultimately becoming enriched at discrete loci termed cohesin-associated regions (Blat and Kleckner, 1999; Laloraya et al., 2000). Cohesion is established during S-phase in a process that requires the Chromosome Transmission Fidelity protein (Ctf7), which is also known as Eco1 (Skibbens et al., 1999; Toth et al., 1999) and involves the replication fork (Kenna and Skibbens, 2003; Lengronne et al., 2006). In budding yeast (Saccharomyces cerevisiae) and fission yeast (Schizosaccharomyces pombe), cohesin complexes remain on the chromosomes until mitotic anaphase (Uhlmann et al., 1999, 2000; Tomonaga et al., 2000). In contrast, in vertebrates, most cohesin complexes are released from the chromosomes during prophase in a separase-independent process (Waizenegger et al., 2000; Losada et al., 2002). The small fraction of cohesin that remains primarily in centromeric regions is released to start anaphase (Sumara et al., 2000). The release of chromosome cohesion at the metaphase-to-anaphase transition is triggered by the Cys protease, separase (ESP1), which specifically cleaves the α-kleisin subunit (Ciosk et al., 1998; Uhlmann et al., 1999, 2000; Buonomo et al., 2000; Hauf et al., 2001). Prior to the metaphase-to-anaphase transition, securin inhibits the protease activity of separase. At the onset of anaphase, securin is degraded by the anaphase-promoting complex/cyclosome freeing separase, which cleaves SCC1, facilitating the release of cohesion and chromosome separation (Cohen-Fix et al., 1996; Ciosk et al., 1998).Studies on the distribution of cohesin proteins during meiosis in a number of organisms, including yeast, Caenorhabditis elegans, mammals, and Arabidopsis (Arabidopsis thaliana), have shown that similar to the situation during mitosis in animal cells, a significant amount of cohesin is either removed from or redistributed on prophase chromosomes in a separase-independent process (Pasierbek et al., 2001; Cai et al., 2003; Eijpe et al., 2003; Lee et al., 2003; Yu and Koshland, 2005). The final resolution of chiasmata, formed as the result of homologous chromosome recombination, and the separation of homologous chromosomes depends on separase cleavage of the meiotic α-kleisin subunit, REC8, along chromosome arms at anaphase I (Buonomo et al., 2000; Kitajima et al., 2003). Centromeric cohesion is protected by the conserved SGO family of proteins until anaphase II when separase cleavage of REC8 facilitates the separation of sister chromatids (Rabitsch et al., 2003; Katis et al., 2004; McGuinness et al., 2005).In addition to its highly conserved role in cleaving the α-kleisin subunit, separase appears to have acquired additional diverse activities in different organisms (Queralt and Uhlmann, 2005). For example, separase plays a role in DNA repair by promoting the redistribution of cohesin complexes to sites of DNA damage during mitotic interphase in budding and fission yeast (Nagao et al., 2004; Strom et al., 2004). Separase is also important for mitotic anaphase spindle assembly and elongation (Jensen et al., 2001; Papi et al., 2005; Baskerville et al., 2008), interphase spindle pole body positioning (Nakamura et al., 2002), and spindle formation during meiosis in yeast (Buonomo et al., 2003). It is also important for the proper positioning of the centrosomes during the first asymmetric mitotic division, eggshell development in C. elegans (Siomos et al., 2001; Rappleye et al., 2002), and for epithelial cell reorganization and dynamics in Drosophila melanogaster (Pandey et al., 2005). In zebra fish, a separase mutation causes genome instability and increased susceptibility to epithelial cancer (Shepard et al., 2007).Results from the characterization of Arabidopsis separase suggested that the protein also has multiple roles in plants (Liu and Makaroff, 2006). Seeds homozygous for a T-DNA insert in Arabidopsis ESP exhibited embryo arrest at the globular stage with the endosperm exhibiting a weak titan-like phenotype. Furthermore, expression of ESP RNA interference (RNAi) from the meiosis-specific DMC1 promoter disrupted the proper removal of the SYN1 cohesin protein from chromosomes during meiosis and resulted in the presence of a mixture of fragmented chromosomes and intact bivalents. The presence of large numbers of intact bivalents led the authors to suggest that in addition to its requirement for the removal of cohesin, ESP may also be required for either the proper attachment of the kinetochores to the spindle or spindle function. These findings, along with the observations that separase appears to have multiple roles in other organisms, led us to conduct a detailed characterization of meiosis in ESP RNAi plants.In this report, we show that meiotic expression of ESP RNAi blocks the release of sister chromatid cohesion during both meiosis I and II, results in nonhomologous centromere association, disrupts the radial microtubule system (RMS) after telophase II, and affects the proper establishment of nuclear cytoplasmic domains. Unlike the large majority of plant meiotic mutants that have been characterized to date, reduction of ESP levels during meiosis leads to the formation of multinucleate microspores.     

Compact Parkin only: insights into the structure of an autoinhibited ubiquitin ligase

EMBO J 32 15, 2099–2112 doi:10.1038/emboj.2013.125; published online May312013Mutations in Parkin represent ∼50% of disease-causing defects in autosomal recessive-juvenile onset Parkinson''s disease (AR-JP). Recently, there have been four structural reports of autoinhibited forms of this RING-IBR-RING (RBR) ubiquitin ligase (E3) by the Gehring, Komander, Johnston and Shaw groups. The important advances from these studies set the stage for the next steps in understanding the molecular basis for Parkinson''s disease (PD).Regulated protein degradation requires that E3s and their access to substrates be exquisitely controlled. RBR family E3s provide striking examples of this regulation. The complex and compact structures of Parkin (Riley et al, 2013; Spratt et al, 2013; Trempe et al, 2013; Wauer and Komander, 2013) as well as another RBR E3, human homologue of Ariadne (HHARI) (Duda et al, 2013), demonstrate extraordinarily intricate inter-domain arrangements. These autoinhibited structures ensure that their functions are restricted until activated.Until recently, RBR E3s were believed to be a subclass of RING E3s, which allosterically activate E2 conjugated with ubiquitin (E2∼Ub). However, Wenzel et al (2011) determined that they are actually hybrid E3s, containing an E2 binding site in RING1 and a catalytic cysteine residue in the domain designated as RING2. The catalytic cysteine is an acceptor for an ubiquitin from RING1-bound E2∼Ub forming an intermediate (E3∼Ub) that leads to substrate or autoubiquitination. In this way, RBRs resemble HECT E3s, which also form catalytic intermediates in ubiquitination. There are 13 human RBR family E3s. Besides Parkin, two notable RBRs are HOIL-1 and HOIP, which form part of a complex integral to NF-κB activation (Wenzel and Klevit, 2012).In addition to causal roles in AR-JP, single allele mutations of Parkin are found in some sporadic cases of PD (references in Wauer and Komander, 2013). Mutations in the Parkin-associated kinase PINK1, which is upstream of Parkin, also account for a significant number of AR-JP cases (Hardy et al, 2009; Narendra et al, 2012; Lazarou et al, 2013). A number of diverse Parkin substrates have been postulated to be associated with PD. There is substantial evidence that one role for Parkin is at mitochondria. Once activated and recruited to damaged/depolarized mitochondria by PINK1, it ubiquitinates exposed mitochondrial proteins leading to both proteasomal degradation and mitophagy (Narendra et al, 2012; Sarraf et al, 2013). Parkin has also been implicated in cell surface signalling and as a tumour suppressor (see references in Wauer and Komander, 2013).Parkin encodes five structured domains, beginning with an N-terminal ubiquitin-like domain (UbLD) and followed by four domains that each bind two zinc (Zn) atoms (Figure 1A). The most N-terminal of the Zn-binding domains is RING0. C-terminal to this is the RBR, consisting of RING1, the IBR and RING2. The crystal structures of inactive Parkin from Riley et al (2013), Trempe et al (2013) and Wauer and Komander (2013) show remarkable congruity. Spatially, the IBR is at the complete opposite end of the molecule from RING2, to which it is connected by a partially unstructured ∼37 residue linker. This linker includes a two-turn helix, referred to as the repressor element of Parkin (REP) or tether, which binds and occludes the E2 binding face of RING1. RING1 occupies the central position in these structures, and RING0 separates RING1 from RING2 (Figure 1B and C). The latter contains the residue identified by Wenzel et al (2011), and confirmed by all three groups, to be the catalytic cysteine, C431. A lower resolution structure also includes the UbLD and places this domain adjacent to RING1 (Trempe et al, 2013). A second unstructured linker connects the UbLD and RING0. UbLDs are involved in a number of protein–protein interactions and small angle X-ray scattering confirms that this domain is integral to the core structure of Parkin (Spratt et al, 2013; Trempe et al, 2013). Biophysical characterization of Parkin and HHARI suggests that each is a monomer in solution.Open in a separate windowFigure 1Schematic and spatial representation of Parkin. (A) Primary structure and domain designations of Parkin, including the REP sequence within the otherwise unstructured IBR-RING2 linker. (B) Structural representation of full-length Parkin (PDB 4K95) highlighting the complex domain interactions in the three-dimensional structure, the catalytic C431 residue, and residue W403 within the REP, which plays a role in stabilizing the autoinhibited form of Parkin. (C) A model of Parkin with the E2 UbcH5B/Ube2D2 bound (devised using PDB 4K95 and PDB 4AP4 to mimic the position of an E2 bound to RING1) to illustrate the required displacement of UbLD and REP and the large distance between the E2∼Ub attachment site of the E2 and the catalytic active site of Parkin. Note that in this conformation the catalytic Cys within RING2 (C431) remains buried by RING0.RING1 is the only bona fide RING domain. All NMR and crystal structures of IBR domains from Parkin, HHARI and HOIP (PDB ID: 2CT7) are in good agreement. The Parkin and HHARI RING2s are structurally highly homologous and share a common Zn-coordinating arrangement with IBR domains. In contrast to the IBR and RING2, RING0 has a distinct arrangement of Zn-coordinating residues (Beasley et al, 2007; Duda et al, 2013; Riley et al, 2013; Spratt et al, 2013; Trempe et al, 2013; Wauer and Komander, 2013) (see Figure 1F of Trempe et al (2013) for the various Zn coordination arrangements).All of the Parkin crystal structures represent inactive forms of the E3. This is imposed by the quaternary positioning of the domains, which precludes activity in multiple ways. RING0 plays two obvious roles to maintain Parkin in an inactive state. RING0 shares an interface with RING2 and buries C431, making it unavailable as an ubiquitin acceptor. Moreover, RING0 intervenes between RING1 and RING2, creating an insurmountable separation of >50 Å between the active site Cys of an E2 bound to RING1 and C431 (Figure 1B and C). Thus, RING0 must be displaced for ubiquitin transfer to occur. Accordingly, deletion of RING0 results in a marked increase in Parkin autoubiquitination and in C431 reactivity (Riley et al, 2013; Trempe et al, 2013; Wauer and Komander, 2013). In HHARI, these two inhibitory functions are fulfilled by the C-terminal Ariadne domain, which similarly interposes between RING1 and RING2 (Duda et al, 2013).Additional inhibition is provided by the REP, which binds to RING1 at the canonical RING-E2 binding site and prevents E2 binding. This provides at least a partial explanation for the impaired ability of Parkin to bind E2 when compared to HHARI, which lacks this element (Duda et al, 2013). A disease-associated REP mutant (A398T) at the RING1 interface increases autoubiquitination (Wauer and Komander, 2013). The significance of inhibition by REP-RING1 binding was verified by mutating a critical RING1-interacting REP residue (W403A). This increased autoubiquitination and E2 binding (Trempe et al, 2013). Consistent with the requirement for charging C431 with ubiquitin in mitochondrial translocation (Lazarou et al, 2013), Parkin association with depolarized mitochondria is accelerated with this mutation (Trempe et al, 2013). Interestingly, W403 also interacts with the C-terminal Val of Parkin within RING2, and could therefore potentially further stabilize the autoinhibited form of the protein (Riley et al, 2013), consistent with previous observations (Henn et al, 2005).The quaternary structure of full-length Parkin also suggests that displacement of its N-terminal UbLD must occur for full activation (Trempe et al, 2013). The positioning of the UbLD adjacent to RING1 indicates that it would provide a steric impediment to E2∼Ub binding (Figure 1B and C). Additionally, displacement of the UbLD could be important to relieve interactions with the IBR-RING2 linker, which, as suggested in a previous study (Chaugule et al, 2011), might help to maintain Parkin in an inactive state. Finally, the crystal structure of the full-length Parkin indicates that the UbLD is not available for interactions with other proteins. This would limit Parkin''s range of intermolecular interactions.RBR E3s have at least two domains critical for sequential ubiquitin transfer and full activity, RING1 and RING2. The RING1 of Parkin, as well as all other RBR E3s, is notable in lacking the basic residue in the second Zn coordinating loop (or its equivalent in U-box proteins), which has recently been implicated in RING-mediated transfer of Ub from E2∼Ub (Metzger et al, 2013). This suggests that other factors play compensatory roles in positioning ubiquitin for transfer from E2∼Ub to C431. A non-mutually exclusive possibility is that the lack of this basic residue in RING1 limits unwanted attack on the E2∼Ub linkage, thereby minimizing the unregulated ubiquitination. Turning to RING2, the area surrounding the active site C431 of Parkin is notable in that it includes a sequence recognizable as a catalytic triad, similar to that in deubiquitinating enzymes. The Cys-His-Glu grouping, found in Parkin and other RBR E3s, contributes to in vitro activity (Riley et al, 2013; Wauer and Komander, 2013). Interestingly, however, the Glu was dispensable in a cellular assay (Riley et al, 2013). This triad is conserved in HHARI, where an Asn between the Cys and His residues (found in a number of RBRs but not conserved in Parkin), was found to be important for catalysis (Duda et al, 2013).The advances made in these studies impart significant information about an important and clinically relevant E3. However, Parkin, as well as HHARI, has been captured in their inactive, unmodified forms. One obvious question is how does Parkin transition between inactive and active states. PINK1 is implicated in phosphorylating Parkin on its UbLD and potentially other sites, with evidence that phosphorylation contributes to Parkin activation (Narendra et al, 2012). How phosphorylation could contribute to protein interactions that might facilitate Parkin activation, potentially including Parkin oligomerization (Lazarou et al, 2013), is unknown. Regardless, it is evident that considerable unwinding of its quaternary structure must take place.While there is much work ahead to understand these processes, one important interface that must be disrupted for activation is that between the REP and RING1. It is intriguing to consider that such interruption might be associated with other alterations in the IBR-RING2 linker, potentially facilitating the movement of the UbLD from RING1 and contributing to activation. Related to activation is the all-important question of how Parkin recognizes and targets specific substrates. While the UbLD represents a potential site of interaction, most purported substrates are not known to have UbLD-interaction domains. Although interactions involving the UbLD could occur indirectly, through bridging molecules, there is also evidence that other regions of Parkin, including the RBR region, might recognize substrates either directly or indirectly (Tsai et al, 2003) and that some substrates may be phosphorylated by PINK1 (Narendra et al, 2012). Conformational changes induced by substrate interactions, particularly in the IBR RING2 linker, could, as above, represent an important aspect of activation.There are over 75 missense mutations of Parkin associated with AR-JP, most of these inactivate the protein, but there are also some that are activating (Wauer and Komander, 2013). Activating mutations presumably result in pathology at least partially as a consequence of increased autoubiquitination and degradation (e.g., A398T). The current studies help to provide a classification of missense mutations into those that affect (i) folding or stability, (ii) catalytic mechanism, and (iii) interactions between domains. Interdomain mutations might inactivate or contribute to constitutive activation leading to autoubiquitination and degradation.Finally, we know little about how the autosomal recessive and the much more prevalent sporadic forms of PD overlap in their molecular pathology. However, mitochondrial dysfunction is increasingly a common theme. Thus, with the structure of the inactive protein in hand, there is hope that we can begin to consider ways in which domain interactions might be altered in a controlled manner to activate, but not hyperactivate, this critical E3 and lessen the progression of PD.     

Split decisions: oesophageal progenitor cell behaviour

Science advance online publication July192012; doi:10.1126/science.1218835The maintenance and regeneration of continually shedding epithelial tissues that make up the linings and barriers of our bodies requires rapid and continual input of proliferative progenitor cells for tissue homeostasis. The mechanisms by which epithelial progenitors cells maintain tissues remain controversial. In a recent Science paper, Doupé et al (2012) demonstrate that a population of equivalent progenitor cells support tissue homeostasis of the oesophagus without the need for slow cycling cells as described in other rapidly dividing epithelia.In tissues such as blood and skin in which differentiated cells constantly turnover, proliferative progenitor populations are required to continually produce lost differentiated cells. Several models have been proposed to explain mechanisms by which progenitor cells contribute to tissue maintenance (Figure 1). A hierarchical model has been suggested in which longer lived stem cells, which may also cycle slowly, produce highly proliferative cells with less self-renewal potential that differentiate into a restricted number of cells. Following proliferative cells in pulse-chase experiments and genetic lineage tracing has supported a hierarchical model in the blood, epidermis and intestine (Fuchs, 2009). Alternatively, an equivalency model has been proposed in which all proliferative progenitor cells are equally able to produce proliferative and differentiated progeny in a stochastic manner. Analysis of labelled clones has supported an equivalency model for progenitors in the interfollicular epidermis and intestine (Clayton et al, 2007; Doupé et al, 2010; Snippert et al, 2010).Open in a separate windowFigure 1Two types of models have been put forward to describe the pattern of progenitor behaviour within mammalian tissues. In the hierarchical model, a stem cell can produce proliferative progenitors with less self-renewal potential that differentiate into lineage-specific cells. Alternatively, an equivalency model has been proposed that assumes equal behaviour of progenitor cells to maintain tissue homeostasis.An elevated interest in understanding the dynamics of oesophageal epithelium has resulted, in part, from the rapid increase in the incidence of oesophageal adenocarcinoma (Devesa et al, 1998). The oesophagus is a stratified epithelium that lacks any appendages or glands, and thus consists of a basal layer of proliferative keratinocytes and several suprabasal layers of differentiated cells, which are continually shed. Previously, labelling of proliferative cells with DNA analogues has demonstrated that proliferation is restricted to the basal cells, which all proliferate in 5 days seemingly stochastically, supporting an equivalency model (Marques-Periera and Leblond, 1965). In contrast, studies using chimeric mice have suggested that proliferation of labelled progenitor cells occurs in a hierarchical manner (Thomas et al, 1988; Croagh et al, 2008).To address this controversy, a recent study in Science uses several genetic mouse models to define the contribution of proliferative basal cells to oesophageal homeostasis (Doupé et al, 2012). In one mouse model, the authors utilized a genetic pulse-chase system based on the tetracycline-regulated expression of the histone H2B-GFP (Tumbar et al, 2004). They find that the rapidly dividing epithelial cells of the oesophagus lose H2B-GFP expression after 4 weeks. These data suggest that either H2B-GFP is degraded (Waghmare et al, 2008) or oesophageal progenitor cells proliferate faster than their counterparts in skin epithelial appendages or blood lineages, which retain H2B-GFP after 4 weeks (Tumbar et al, 2004; Foudi et al, 2009).To analyse the properties of oesophageal progenitor cells in more detail, the authors label single cells using an inducible cre-lox genetic system and followed clones for a year. Similar to their results with this system in the tail and ear epidermis (Clayton et al, 2007; Doupé et al, 2010), the authors find that the size of the persistent clones is linear with time. Statistical analysis of the clone size data supports the ability of the cells to contribute to proliferative and non-proliferative (i.e., differentiated) progeny with equal probability. Thus, these data support a model in which all of the labelled cells are equivalent.In addition to homeostasis, the authors explore how proliferative progenitors contribute to alterations in tissue homeostasis. After inflicting wounds by biopsy, marked clones span both proliferative and non-proliferative zones of the healing oesophageal epithelium, suggesting that they maintain a progenitor fate with distinct phenotypes. With atRA treatment, the authors show that suprabasal cell formation increases, which is consistent with the known effect of atRA on the oesophagus (Lasnitzki, 1963). Statistical analysis reveals that the probability of forming basal and suprabasal cells was not altered with atRA administration. However, since proliferative cells exist in suprabasal layers during epithelial hyperplasia, additional analyses of cell state are required to determine if atRA maintains stochastic fate decisions of progenitor cells. Furthermore, the progenitor response to atRA treatment might be limited by niche space along the basement membrane like in intestinal crypt progenitor cells (Snippert et al, 2010).In summary, this study together with the authors'' previous work provides additional support for the existence of equivalent progenitor cells within stratified epithelium in several tissues. Additional studies revealing how epithelial progenitor cells behave when proliferation and differentiation are altered in the oesophagus could shed light on mechanisms for the pathogenesis of oesophageal tumours or diseases such as Barrett''s oesophagus.     

A structural road map to unveil basal body composition and assembly

EMBO J 31 3, 552–562 (2012); published online December132011The Basal Body (BB) acts as the template for the axoneme, the microtubule-based structure of cilia and flagella. Although several proteins were recently implicated in both centriole and BB assembly and function, their molecular mechanisms are still poorly characterized. In this issue of The EMBO journal, Li and coworkers describe for the first time the near-native structure of the BB at 33 Å resolution obtained by Cryo-Electron Microscopy analysis of wild-type (WT) isolated Chlamydomonas BBs. They identified several uncharacterized non-tubulin structures and variations along the length of the BB, which likely reflect the binding and function of numerous macromolecular complexes. These complexes are expected to define BB intrinsic properties, such as its characteristic structure and stability. Similarly to the high-resolution structures of ribosome and nuclear pore complexes, this study will undoubtedly contribute towards the future analysis of centriole and BB biogenesis, maintenance and function.The microtubule (MT)-based structure of the cilium/flagellum grows from the distal part of the Basal Body (BB), which in many animal cells develops from the mature centriole in the centrosome. Electron microscopic (EM) images of chemically fixed resin-embedded centrioles and basal bodies (CBBs) suggest that their ultrastructure is similar, and that their key components are MTs. The mechanisms underlying the organization of CBB MTs, comprising highly stable closed and open MTs, are likely to hold many surprises as they are remarkably different from other microtubular structures in the cell. Additionally, non-MT-based structures are also part of the CBB, including a cartwheel in the proximal lumen region that reinforces CBB symmetry (reviewed in Azimzadeh and Marshall, 2010 and Carvalho-Santos et al, 2011).Several centriole components and BB proteins were identified by comparative and/or functional genomics and proteomics studies of purified CBBs (reviewed in Azimzadeh and Marshall, 2010 and Carvalho-Santos et al, 2011). Advances in our understanding of the molecular mechanisms of CBB assembly depend on high-resolution comparative studies of wild-type (WT) and mutant structures, as well as characterization of the localization of molecular complexes within the small CBB structure. Despite the existence of beautiful ultrastructure data acquired from chemically fixed specimens (Geimer and Melkonian, 2004; Ibrahim et al, 2009), high-resolution structures of native CBBs were missing. Using electron cryo-tomography and 3D subtomogram averaging, Li et al (2012) solved the structure of the near-native BB triplet at 33 Å resolution. A pseudo-atomic model of the tubulin protofilaments at the core of the triplets was built by fitting the atomic structure of α/β-tubulin monomers into the BB tomograms.The 3D density map reveals several additional densities that represent non-tubulin proteins attached, both internally and externally, to all triplet MTs, some linking MTs inside the triplets and/or MTs in consecutive triplets (Li et al, 2012; for a summary, see Li et al, 2012; Geimer and Melkonian, 2004; Ibrahim et al, 2009), but with less detail and complexity. The authors speculate that some of the additional densities present at the A- and B-tubule inner wall might correspond to proteins of the tektin family, probably conferring rigidity to the BB triplet (Amos, 2008).

Table 1

Characteristics of the non-α/β-tubulin structures reported in Li et al (2012) in this issue of The EMBO journal

Enlightening the impact of immunogenic cell death in photodynamic cancer therapy

EMBO J 31 5, 1062–1079 (2012); published online January172012In this issue of The EMBO Journal, Garg et al (2012) delineate a signalling pathway that leads to calreticulin (CRT) exposure and ATP release by cancer cells that succumb to photodynamic therapy (PTD), thereby providing fresh insights into the molecular regulation of immunogenic cell death (ICD).The textbook notion that apoptosis would always take place unrecognized by the immune system has recently been invalidated (Zitvogel et al, 2010; Galluzzi et al, 2012). Thus, in specific circumstances (in particular in response to anthracyclines, oxaliplatin, and γ irradiation), cancer cells can enter a lethal stress pathway linked to the emission of a spatiotemporally defined combination of signals that is decoded by the immune system to activate tumour-specific immune responses (Zitvogel et al, 2010). These signals include the pre-apoptotic exposure of intracellular proteins such as the endoplasmic reticulum (ER) chaperon CRT and the heat-shock protein HSP90 at the cell surface, the pre-apoptotic secretion of ATP, and the post-apoptotic release of the nuclear protein HMGB1 (Zitvogel et al, 2010). Together, these processes (and perhaps others) constitute the molecular determinants of ICD.In this issue of The EMBO Journal, Garg et al (2012) add hypericin-based PTD (Hyp-PTD) to the list of bona fide ICD inducers and convincingly link Hyp-PTD-elicited ICD to the functional activation of the immune system. Moreover, Garg et al (2012) demonstrate that Hyp-PDT stimulates ICD via signalling pathways that overlap with—but are not identical to—those elicited by anthracyclines, which constitute the first ICD inducers to be characterized (Casares et al, 2005; Zappasodi et al, 2010; Fucikova et al, 2011).Intrigued by the fact that the ER stress response is required for anthracycline-induced ICD (Panaretakis et al, 2009), Garg et al (2012) decided to investigate the immunogenicity of Hyp-PDT (which selectively targets the ER). Hyp-PDT potently stimulated CRT exposure and ATP release in human bladder carcinoma T24 cells. As a result, T24 cells exposed to Hyp-PDT (but not untreated cells) were engulfed by Mf4/4 macrophages and human dendritic cells (DCs), the most important antigen-presenting cells in antitumour immunity. Similarly, murine colon carcinoma CT26 cells succumbing to Hyp-PDT (but not cells dying in response to the unspecific ER stressor tunicamycin) were preferentially phagocytosed by murine JAWSII DCs, and efficiently immunized syngenic BALB/c mice against a subsequent challenge with living cells of the same type. Of note, contrarily to T24 cells treated with lipopolysaccharide (LPS) or dying from accidental necrosis, T24 cells exposed to Hyp-PDT activated DCs while eliciting a peculiar functional profile, featuring high levels of NO production and absent secretion of immunosuppressive interleukin-10 (IL-10) (Garg et al, 2012). Moreover upon co-culture with Hyp-PDT-treated T24 cells, human DCs were found to secrete high levels of IL-1β, a cytokine that is required for the adequate polarization of interferon γ (IFNγ)-producing antineoplastic CD8⁺ T cells (Aymeric et al, 2010). Taken together, these data demonstrate that Hyp-PDT induces bona fide ICD, eliciting an antitumour immune response.By combining pharmacological and genetic approaches, Garg et al (2012) then investigated the molecular cascades that are required for Hyp-PDT-induced CRT exposure and ATP release. They found that CRT exposure triggered by Hyp-PDT requires reactive oxygen species (as demonstrated with the ¹O₂ quencher L-histidine), class I phosphoinositide-3-kinase (PI3K) activity (as shown with the chemical inhibitor wortmannin and the RNAi-mediated depletion of the catalytic PI3K subunit p110), the actin cytoskeleton (as proven with the actin inhibitor latrunculin B), the ER-to-Golgi anterograde transport (as shown using brefeldin A), the ER stress-associated kinase PERK, the pro-apoptotic molecules BAX and BAK as well as the CRT cell surface receptor CD91 (as demonstrated by their knockout or RNAi-mediated depletion). However, there were differences in the signalling pathways leading to CRT exposure in response to anthracyclines (Panaretakis et al, 2009) and Hyp-PDT (Garg et al, 2012). In contrast to the former, the latter was not accompanied by the exposure of the ER chaperon ERp57, and did not require eIF2α phosphorylation (as shown with non-phosphorylatable eIF2α mutants), caspase-8 activity (as shown with the pan-caspase blocker Z-VAD.fmk, upon overexpression of the viral caspase inhibitor CrmA and following the RNAi-mediated depletion of caspase-8), and increased cytosolic Ca²⁺ concentrations (as proven with cytosolic Ca²⁺ chelators and overexpression of the ER Ca²⁺ pump SERCA). Moreover, Hyp-PDT induced the translocation of CRT at the cell surface irrespective of retrograde transport (as demonstrated with the microtubular poison nocodazole) and lipid rafts (as demonstrated with the cholesterol-depleting agent methyl-β-cyclodextrine). Of note, ATP secretion in response to Hyp-PDT depended on the ER-to-Golgi anterograde transport, PI3K and PERK activity (presumably due to their role in the regulation of secretory pathways), but did not require BAX and BAK (Garg et al, 2012). Since PERK can stimulate autophagy in the context of ER stress (Kroemer et al, 2010), it is tempting to speculate that autophagy is involved in Hyp-PDT-elicited ATP secretion, as this appears to be to the case during anthracycline-induced ICD (Michaud et al, 2011).Altogether, the intriguing report by Garg et al (2012) demonstrates that the stress signalling pathways leading to ICD depend—at least in part—on the initiating stimulus (Figure 1). Speculatively, this points to the coexistence of a ‘core'' ICD signalling pathway (which would be common to several, if not all, ICD inducers) with ‘private'' molecular cascades (which would be activated in a stimulus-dependent fashion). Irrespective of these details, the work by Garg et al (2012) further underscores the importance of anticancer immune responses elicited by established and experimental therapies.Open in a separate windowFigure 1Molecular mechanisms of immunogenic cell death (ICD). At least three processes underlie the immunogenicity of cell death: the pre-apoptotic exposure of calreticulin (CRT) at the cell surface, the secretion of ATP, and the post-apoptotic release of HMGB1. ICD can be triggered by multiple stimuli, including photodynamic therapy, anthracycline-based chemotherapy, and some types of radiotherapy. The signalling pathways elicited by distinct ICD inducers overlap, but are not identical. In red are indicated molecules and processes that—according to current knowledge—may be required for CRT exposure and ATP secretion in response to most, if not all, ICD inducers. The molecular determinants of the immunogenic release of HMGB1 remain poorly understood. ER, endoplasmic reticulum; P-eIF2α, phosphorylated eIF2α; PI3K, class I phosphoinositide-3-kinase; ROS, reactive oxygen species.     

Dynein at the nuclear envelope

Most cellular organelles are positioned through active transport by motor proteins. The authors discuss the evidence that dynein has important cell cycle-regulated functions in this context at the nuclear envelope.Most cellular organelles are positioned through active transport by motor proteins. This is especially important during cell division, a time when the organelles and genetic content need to be divided equally between the two daughter cells. Although individual proteins can attain their correct location by diffusion, larger structures are usually positioned through active transport by motor proteins. The main motor that transports cargoes to the minus ends of the microtubules is dynein. In nondividing cells, dynein probably transports or positions the nucleus inside the cells by binding to the nuclear envelope (NE; Burke & Roux, 2009). However, it appears that dynein also has important cell-cycle-regulated functions at the NE, as it is recruited to the NE every cell cycle just before cells enter mitosis (Salina et al, 2002; Splinter et al, 2010). Here, we discuss why dynein might be recruited to the NE for a brief period before mitosis.During late G2 or prophase the centrosomes separate to opposite sides of the nucleus, but remain closely associated with the NE during separation. This close association is probably mediated through NE-bound dynein, which ‘walks'' towards the minus ends of centrosomal microtubules, thereby pulling centrosomes towards the NE (Splinter et al, 2010; Gonczy et al, 1999; Robinson et al, 1999). We speculate that close association of centrosomes to the NE might have several functions. First, if centrosomes are not mechanically coupled to the NE, centrosome movement during separation will occur in random directions and chromosomes will not end up between the two separated centrosomes. In this scenario, individual kinetochores might attach more frequently to microtubules coming from both centrosomes (merotelic attachments), a defect that can result in aneuploidy, a characteristic of cancer. Second, centrosome-nuclear attachment also keeps centrosomes in close proximity to chromosomes, which might facilitate rapid capture of chromosomes by microtubules nucleated by the centrosomes after NE breakdown. This might not be absolutely essential, as chromosome alignment can occur in the absence of centrosomes. However, the spatial proximity of centrosomes and chromosomes at NE breakdown might improve the fidelity of kinetochore capture and chromosome alignment.In addition, dynein has also been suggested to promote centrosome separation in prophase in some systems (Gonczy et al, 1999; Robinson et al, 1999; Vaisberg et al, 1993), although not in others (Tanenbaum et al, 2008). Perhaps dynein, anchored at the NE just before mitosis, could exert force on microtubules emanating from both centrosomes, thereby pulling centrosomes apart. However, this force could also be produced by cortical dynein and specific inhibition of NE-associated or cortical dynein will be required to test which pool is responsible.Dynein has also been implicated in the process of NE breakdown itself, by promoting mechanical shearing of the NE. Two elegant studies showed that microtubules can tear the NE as cells enter mitosis (Salina et al, 2002; Beaudouin et al, 2002). One possibility is that microtubules growing into the NE mechanically disrupt it. Alternatively, NE-associated dynein might ‘walk'' along centrosomal microtubules and thereby pull on the NE, tearing it apart. However, testing the exact role of dynein in NE breakdown is complicated by the fact that centrosomes detach from the NE on inactivation of dynein and centrosomal microtubules stop growing efficiently into the NE. Thus, selective inhibition of dynein function will also be required to test this idea.Specific recruitment of dynein to the NE just before mitosis clearly suggests a role for dynein at the NE in preparing cells for mitosis. A major role of NE-associated dynein is to maintain close association of centrosomes with the NE during centrosome separation, which might be needed for efficient capture and alignment of chromosomes after NE breakdown, but additionally, NE-associated dynein could facilitate breakdown and contribute to centrosome separation in some systems.