Encounters with transcription in a six-day journey through the nucleus

Abstract

An outstanding array of renowned scientists gathered at this year's Cold Spring Harbor Symposium on Quantitative Biology on “nuclear Organization & Function” to discuss a variety of processes that take place inside the nucleus. Highlighted in this report are the talks related to the regulation of transcriptional processes.

The LXXV Cold Spring Harbor Symposium on Quantitative Biology held June 2–7, 2010 focused on aspects of “Nuclear Organization & Function.” Organized by Terri Grodzicker, David Spector, David Stewart and Bruce Stillman, the meeting brought together an outstanding array of renowned scientists that presented and discussed with colleagues their recent findings. Speakers covered a vast variety of nuclear processes: from DNA replication and genome integrity to chromosome structure and mitosis, aspects of RNA regulation and events that occur in the nuclear periphery. This report highlights only selected emerging themes that are pertinent in one way or another to the process of transcription.

It has been known for quite some time that transcription, and the production of small non-coding RNAs, are essential to the formation of heterochromatic structures. Heterochromatin not only plays crucial structural roles, but is also critical for gene regulation, centromere formation and function, recombination, transcriptional and prost-transcriptional silencing of repetitive DNA elements and sister-chromatid cohesion. Several speakers discussed recent progress in the understanding of the roles of transcription and RNA molecules in heterochromatin formation and function. Robin Allshire, from the University of Edinburgh in the UK, talked about heterochromatin function at the centromere. In fission yeast, centromeres consist of a central domain containing the histone H3 variant CENP-A, flanked by heterochromatic repeats. Allshire's lab is interested in understanding the features that promote de novo CENP-A assembly at the centromere central domain. He and colleagues have shown that heterochromatin formation at the outer repeats (and the consequent CENP-A incorporation at the central domain) was dependent on RNAi, the histone H3 lysine 9 (H3K9) methyltransferase Clr4 and Swi6/HP1. They now demonstrated that artificial tethering of Clr4 upstream of a reporter gene was sufficient to silence gene expression in a locus-independent manner and independently of the RNAi machinery, which indicates that Clr4 bypasses the need for centromeric outer repeats and RNAi. Also intrigued by heterochromatin, Rob Martienssen, from Cold Spring Harbor Laboratory in NY, is interested in understanding how RNA polymerase II is able to transcribe heterochromatin, which, by definition, is transcriptionally repressed. Martienssen eloquently explained that the answer lies in the cell cycle. His group found that in fission yeast transcription of centromeric repeats occurs specifically during the S phase, at a time when Swi6 is partially displaced from centromeric regions due to the phosphorylation of histone H3 serine 10 (H3S10), which interferes with the binding of Swi6 to H3K9me. It is at this time that RNA polymerase II produces both forward and reverse centromeric transcripts that accumulate transiently and allow the formation of siRNAs. RNAi then mediates the recruitment of histone modifiers at the centromere to establish heterochromatin, mediating the inheritance of heterochromatic modifications. RNAi-induced heterochromatin formation is therefore coupled to the cell cycle, and transcription of heterochromatic regions is limited to a specific window of the cell cycle (the S phase), when transcriptional repression is weakened.

Continuing the discussion of the role of RNAi in heterochromatin programming, Gary Felsenfeld, from the National Institutes of Health in Bethesda, MD, showed that the 16 kb heterochromatic structure upstream of the chicken HS4 insulator, located 5′ of the β-globin gene, depends on Dicer. The chicken homolog of Ago2 was also found to bind to the 16 kb heterochromatic domain in a Dicer-dependent manner and was necessary for the maintenance of a condensed heterochromatic structure. Interested in the involvement of Ago2 and RNAi in the maintenance of heterochromatic domains in other species, they also analyzed the distribution of Ago2 in human repeat sequences using the human cell line K562. They found that hAgo2 binding sites were enriched in satellite and rRNA repeats, where the ratio of ‘sense’ and ‘antisense’ transcripts was close to 1. Therefore, heterochromatic structures appear to be maintained by conserved mechanisms among vertebrates.

Nicholas Proudfoot, from the University of Oxford in UK, discussed a newly identified class of transient heterochromatin that forms across convergent gene pairs in fission yeast. Monika Gullerova with Proudfoot found that, in G1 phase of the cell cycle, convergent genes did not terminate transcription at the end of the gene. Instead, these genes run through the downstream gene and, consequently, ended up generating dsRNAs. These dsRNAs were recognized by the RNAi machinery and led to the transient formation of heterochromatin across convergent genes. They also observed that cohesin was loaded at this time. Then, in S phase, heterochromatin disappeared, but cohesin remained. In G2 (which is the longest phase of the cell cycle in fission yeast), RNA polymerase II collided with cohesin, no longer running into the downstream gene, terminating transcription. Proudfoot explained how, at the end of G2, cohesin is broken down, dsRNA is made again and the cycle repeats. Interestingly, when investigating what kinds of genes are convergent in fission yeast, Gullerova and Proudfoot found that RNAi components were predominantly encoded by convergent genes. Similar to Martienssen and Grewal's independent observation about centromeric repeats transcription occurring specifically during S phase, they also observed that these genes were downregulated in G1-S phase. Intrigued by this observation, they were curious to understand whether downregulation of convergent RNAi genes was linked to centromeric transcription in S phase. Proudfoot described that by tandemizing convergent RNAi genes, G1 downregulation was lost (and, concomitantly, G1-S centromeric transcription was also reduced), which indicates that the production of read-through transcripts in G1-S is dictated by gene orientation. Heterochromatic marks were also lost over tandemized mutants. In addition, tandemly oriented RNAi mutants were phenotypically smaller and showed mitotic defects. Proudfoot therefore concluded that RNAi convergent gene autoregulation during the cell cycle is necessary for correct centromeric heterochromatin formation and for mitosis.

Moving the discussion to plants, Craig Pikaard, from Indiana University in Bloomington, IN, talked about two plant specific DNA-dependent RNA polymerases, RNA polymerases IV and V, which are specialized forms of RNA polymerase II that play important roles in heterochromatin formation. These polymerases are in charge of producing the precursors of siRNA that direct siRNA-dependent DNA methylation. Pikaard's lab is interested in identifying RNA polymerase IV and V transcripts. Recently, they found that RNA polymerase IV and RNA- dependent RNA polymerase II (RDR2) coexist in the same multifunctional transcriptional complex. In addition, although no RNA polymerase IV transcripts have been detected using any of the classical in vivo assays, Pikaard and colleagues were able to detect transcription in vitro using templates that mimic paused transcription elongation complexes. They also discovered polymerase V-dependent intergenic non-coding transcripts that are required for repressive chromatin modification in cis. These transcripts could be cross-linked to Ago4 and served as scaffold for the binding of Ago4-siRNA complexes. In these ways, RNA polymerases IV and V contribute to large-scale heterochromatin organization.

Other highlights of the meeting were presented on aspects of cell reprogramming, pluripotency and differentiation. Rudolf Jaenisch, from the Whitehead Institute for Biomedical Research in Cambridge, MA, and Huck Hui Ng, from the Genome Institute of Singapore, dissected the molecular mechanisms of pluripotent embryonic stem cells (ESCs). Jaenisch highlighted studies from his lab analyzing the different characteristics of mouse and human pluripotent cells. Conventional human ESCs (hESC) present biological characteristics that stably affect their pluripotency state. In addition, they grow slowly, have low single cloning efficiency and show low efficiency of homologous recombination, which interferes with their ability to be genetically manipulated. With the goal of generating hESCs with mouse ESC (mESC)-like characteristics, Jaenisch and colleagues analyzed whether some of the differences between the two could be attributed to cell culture and isolation conditions. Consistent with culture conditions profoundly affecting the biology of hESCs was the finding that hESCs isolated and propagated under 5% O₂ displayed a pre-X inactivation status (XaXa) and, similar to mESCs, initiated random inactivation upon differentiation. Shifting the cultures to 20% oxygen induced irreversible X chromosome inactivation as is seen in conventional hESC cells. However, X chromosome inactivation could be reversed by propagation of conventional hESCs in LIF/Stat3 and “2i” conditions (the addition of PD/CH inhibitors) and overexpression of Oct4 and Klf4. The converted hESCs corresponded to “naïve” mESCs by multiple criteria: they had reactivated the inactive X chromosome resulting in a pre-X inactivation status (XaXa in female cells), exhibited high single cell cloning efficiency, were dependent on LIF/STAT signaling instead of FGF/Activin, could routinely be passaged in trypsin without suffering chromosomal aberrations and showed a gene expression pattern that was different from conventional hESCs and resembled that of naïve mESCs. When human fibroblasts were reprogrammed under those culture conditions, iPSCs with similar properties were generated. Jaenisch indicated that the next step is to analyze whether these more “naïve” hESCs show higher differentiation potential and facilitate gene targeting. Ng, who is interested in understanding the mechanisms by which transcription factors maintain and specify stem cell identity, talked about a study in which human and mouse ES cells were also compared. Ng and colleagues asked whether Oct4 and Nanog circuitries are conserved between the two species. They found that, even though mouse and humans have conserved Sox2 expression patterns, the binding profiles of these two factors at the Sox2 locus are similar in only 15% of the cases. Ng explained that only 5% of Oct4 and Nanog binding sites found in humans are also found in the mouse. He showed that species-specific transposable elements have contributed to up to 25% of the new sites by rewiring the core regulatory network of ES cells and altering the circuitry that controls pluripotency in ES cells.

Continuing with the topic of pluripotency, Ken Zaret, from the University of Pennsylvania in Philadelphia, focused his talk on the mechanisms by which the embryonic endoderm becomes competent to activate the genetic programs for liver and pancreas cells. Zaret talked about FoxA proteins, which are transcription factors that resemble linker histones in structure and bind nucleosomal DNA early in development, establishing competence for specific activation states. Using fluorescence recovery after photobleaching (FRAP), Zaret and colleagues found that FoxA moved slower than other transcription factors within the nucleus. By analyzing different FoxA mutants, they determined that the slow mobility of FoxA is determined by its ability to bind non-specifically to nucleosomes. Zaret proposed that the differential ability of FoxA to access nucleosomal DNA, something that many other transcription factors are not able to do, and its slow mobility reflect its role as a pioneer factor, establishing competence for expression states early in development. Tom Misteli, from the National Cancer Institute in Bethesda, MD, talked about a different type of stem cells called cancer stem cells (CSCs). CSCs are a distinct subset of cells within a tumor that are capable of initiating and sustaining the growth of the tumor. The CSC model states that only these cells can lead to the formation of new tumors and, therefore, cancer therapies specifically targeted to CSCs have the potential of being extremely successful and specific. Misteli described a system to generate CSCs in vitro by reprogramming of somatic cells using defined factors. Using this system Misteli's lab identified the cell surface marker SSEA1 (stage specific embryonic antigen 1) as a marker of tumorigenic transformed cells. SSEA1⁺ cells showed increased expression of tumor genes and decreased expression of tumor suppressor genes. Whereas SSEA1⁺ cells were able to give rise to the bulk of the tumor mass in vivo, giving rise to heterogeneous lineages of cancer cells, SSEA1⁻ cells were not tumorigenic. Misteli was therefore able to reprogram somatic cells into CSCs that can give rise to cancer, supporting the CSC model.

The section devoted to epigenetic modifications and gene expression brought together enthusiasts of histone variants, insulators, Polycomb proteins, and an interesting newer player: multifunctional PARP. David Allis, from The Rockefeller University in NY, chose to discuss histone variant H3.3. He described work from his lab in which they used genome editing with zinc finger nucleases to tag endogenous H3.3. Allis and colleagues analyzed the H3 variant in ES cells and discovered that, whereas the H3.3 chaperone Hira was required for H3.3 enrichment at active housekeeping genes, this chaperone was not essential for the localization of H3.3 at transcription factor binding sites. Unexpectedly, they identified Atrx and Daxx as H3.3 unique binding partners. Atrx, a SNF2-related chromatin remodeler and Daxx, a death-domain-associated protein, bound to H3.3 even in the absence of Hira. Interestingly, whereas Atrx was required for telomeric deposition of H3.3, Hira was not. In genome-wide ChIP-seq analyses using H3.3 tagged lines that were null for either Hira or Atrx, they showed that Hira and Atrx localized H3.3 to distinct non-overlapping genomic locations, suggesting that distinct factors mediate H3.3 localization at specific genomic locations. Talking now about histone variant dynamics, Steve Henikoff, from the Fred Hutchinson Cancer Research Center in Seattle, WA, explained that the rapidity of histone turnover, which is faster than a cell cycle, suggests that chromatin marks are not themselves inherited. Henikoff and colleagues measured nucleosome turnover dynamics and found it to be most rapid over active genes, sites for trithorax-group protein binding and at replication origins. Contrarily, nucleosome turnover was slowest at polycomb-group protein binding sites. Henikoff suggested that the quantitative differences in turnover rates may be what distinguish active from silent states. In this study, Henikoff and colleagues analyzed the global genetic patterns of H3.3 incorporation in S2 cells over 9,300 genes and observed that, in most cases, H3.3 was low over the promoter region, went up over the gene body, and then went down again towards the 3′ end of the gene, a pattern that coincides with the histone H3 and H4 turnover pattern.

Danny Reinberg, from the HHMI at NYU School of Medicine in NY, talked about the requirement of Polycomb Repressive Complex (PRC2) for the proper maintenance and propagation of the H3K27me3 mark. He described the involvement of a series of steps that include the binding of PRC2 subunit Eed to H3K27me3 and other trimethylated lysines associated with chromatin repression. Reinberg explained that, even though the binding of Eed stimulates the Ezh2 enzymatic activity of the complex, it is not required for PRC2 recruitment. Looking at the factor responsible for PRC2 access to chromatin, Reinberg and colleagues identified Jarid2. Jarid2 contains a JmjC domain but shows no demethylation activity. Instead, Jarid2 was found to interact with PRC2 and stimulate its H3K27 methylation activity. Jarid2 facilitates PRC2 recruitment to target genes and is required for PRC2-repressive activity in ES cells. He suggested that the recruitment of PRC2 to target genes occurs through at least four steps that in isolation are each low energy but that together synergize.

John Lis, from Cornell University in Ithaca, NY, described studies using both live cell optical analyses and biochemical approaches to understand the mechanisms of activation of heat shock genes and the associated changes in chromatin structure. Using the Hsp70 gene, which provides a well-characterized example of promoter-proximal paused RNA polymerase II, he analyzed the role of the specific factors that modulate the response of RNA polymerase II upon heat shock activation. Lis and colleagues have shown that Heat Shock Factor (HSF) was nucleoplasmic before heat shock and was rapidly recruited to Heat Shock (HS) loci, where it was detected within 10 seconds of heat shock, reaching saturation at 2 minutes. Activated HSF was stably bound to HS loci and its turnover was not required for rounds of Hsp70 transcription. Recruitment of HSF was followed by the ordered recruitment of RNA polymerase II, pTEFb, Spt6 and Topo1; each factor being recruited at its own timing or with its own rate, and not as part of a large complex. Interested in identifying the factors required for the initial nucleosome loss that occurs after heat shock, they used a biochemical approach based on an assay called nucleosome scanning to analyze the presence of nucleosomes at different times after heat shock. They found that PARP [Poly (ADP-ribose) Polymerase] was required for the initial nucleosome loss. Before heat shock, PARP was found at the first nucleosome. After heat shock, PARP was lost from the first nucleosome and redistributed along the Hsp70 gene, with kinetics that resembled the kinetics of nucleosome loss. Lis proposed a model in which after heat shock (and recruitment of HSF) PARP mediates the rapid redistribution of nucleosomes by adding PAR chains along the gene. These PAR chains would facilitate the stripping of his- tones from the DNA, which would cause nucleosome loss. Using FRAP, Lis also showed that transcriptional activation by heat shock caused the entry and accumulation of RNA polymerase II, creating a local compartment where the polymerase is efficiently recycled. Importantly, this transcription compartment depends on PARP enzymatic activity.

The discussion on the role of insulators in the protection of genomic regions against the spreading of heterochromatin came by the hand of Felsenfeld. He described his studies of the multi-component functional 5′ insulator element HS4 from the chicken β-globin locus. Insulators are able to block the action of an enhancer on a promoter when positioned in between the two. In addition, insulators also work as a barrier that protects against silencing by adjacent condensed chromatin. Remarkably, the enhancer-blocking and barrier activities appear to have different underlying mechanisms. Felsenfeld discussed studies from his lab analyzing HS4 barrier insulation function. HS4 contains five elements, named FI-V, that are required for its insulator function. FII was previously shown to bind CTCF and is necessary and sufficient for enhancer-blocking; remarkably, it is dispensable for the barrier activity. On the other hand, FI, FIII, FIV and FV are all required for protection against heterochromatin spreading. Felsenfeld and colleagues found that FI, FIII and FV were bound by BGP1/Vezf1, and were responsible for protection against DNA methylation. Interested on how general this activity was, Felsenfeld analyzed a mouse embryonic stem cell line that was null for Vezf1. In these cells, DNA methylation levels were lower in many sites of the genome, and the levels of the de novo methyltransferase Dnmt3b were substantially reduced. They showed that Vezf1 bound sites in an intronic region of Dnmt3b and caused RNA polymerase II pausing, likely affecting the alternative splicing of the enzyme.

RNA polymerase II, alternative splicing and silencing chromatin marks found each other again in the same sentence several times during a section cleverly named “transcription meets RNA processing.” Stephen Buratowski, from Harvard Medical School in Boston, MA, talked about the connections between transcription, chromatin modifications and mRNA processing, stressing the need for an integration of the study of these processes, which are clearly interconnected and regulate each other. He focused his talk on studies from his lab analyzing the roles of the co-transcriptional methylation of H3K4 (mediated by Set1) and H3K36 (mediated by Set2) in the regulation of mRNA processing. H3K4me3 is found at the 5′end of genes and recruits factors associated with the activation of the promoter. In addition, H3K4me2 recruits the Set3 complex, leading to the deacetylation of the 5′ region of the gene. Buratowski and colleagues became interested in the role of H3K4 methylation in early transcriptional termination. In yeast, two types of transcriptional termination pathways exist: the classical mRNA late pathway and the early termination pathway, which functions at snoRNAs and cryptic unstable transcripts (CUTs) and is mediated by the targeting of the Sen1/Nrd1/Nab3 complex to the 5′end of genes. Given that both Sen1/Nrd1/Nab3 complex and H3K4 methylation by Set1 interact with RNA polymerase II CTD Ser5P, Buratowski and colleagues analyzed the genetic interactions between H3K4 methylation and the early termination pathway. They found that loss of Set1 (and, consequently, of H3K4me3) exacerbated the early termination defects of nrd1 mutants and therefore concluded that H3K4 methylation (and the resulting H3 acetylation/deacetylation) affects early elongation, likely allowing more time for the phosphorylation of CTD Ser5 and for the loading of Sen1/Nrd1/Nab3. Moving on to higher eukaryotes, Buratowski proposed the idea that some of the short, unstable, promoter-associated transcripts that are frequently explained by the existence of “paused” polymerases are actually early terminating transcripts. In support of his idea, he explained that the negative elongation factor NELF, which was described to control “pausing,” shares a number of features with the Sen1/Nrd1/Nab3 complex, such as its localization to the 5′end of mRNA, an RNA binding subunit, its induction by CTD Ser5P, and its requirement for the termination of short, non-polyadenylated RNAs.

Alberto Kornblihtt, from the University of Buenos Aires in Argentina, focused his talk on the kinetic coupling of RNA polymerase II and alternative splicing. His lab had confirmed the role of RNA polymerase II elongation rate in alternative splicing with experiments in which they showed that the use of a slow RNA polymerase increased the inclusion of an alternative exon of the fibronectin gene. Intrigued by the mechanism by which the rate of RNA polymerase II elongation could be affected and, in turn, regulate alternative splicing, Kornblihtt and colleagues used siR-NAs targeting pre-mRNA intronic regions downstream of alternative exons in order to generate epigenetic silencing marks. They found that the presence of these marks promoted the inclusion of the alternative exon and that the effect was reversed by treatment with “chromatin opening” agents such as inhibitors of histone deacetylases or histone methyltransferases. Interestingly, depletion of Ago1 or Dicer affects alternative splicing of many genes, and suggests that small RNAs might be involved in this pathway. In a preliminary ChIP-seq genome-wide analysis, Kornblihtt and colleagues identified convergence of silencing marks (H3K27me3, H3K9me2 and H3K36me3) in Ago1 target regions. These data will allow the identification of genomic targets that could be used to test the effect of Ago1-mediated heterochromatin formation and alternative splicing.

Christian Muchardt, from the Institut Pasteur in Paris, France, focused his talk on the role of silencing chromatin marks in the control of alternative splicing. He showed studies using the CD44 gene, which codes for a membrane protein involved in cell adhesion. CD44 can potentially generate more than a thousand isoforms and is highly dependent on the chromatin remodeling factor Brm for activation. Muchardt and colleagues observed that RNA polymerase II, phosphorylated in the CTD at Serine 5, accumulated in the variant region of CD44. Accumulation of RNA polymerase is considered as an indication of a slowed down enzyme. By ChIP walking of the CD44 gene they found that H3K9me3 and H3K9me2 also accumulated in the variant region of CD44 and, more importantly, their removal caused decreased variant exon inclusion. Muchardt also found that the inclusion of CD44 variant exons was a function of HP1γ expression levels. Upon addition of PMA, PKC is activated, resulting in the phosphorylation of HP1γ, which then accumulates on CD44 variant exons. Depletion of HP1γ caused the reduction of RNA polymerase II accumulation and a similar reduction on the accumulation of the splicing factor U2AF65 on the same region of CD44. Interestingly, in a genome-wide study using Affymetrix exon arrays after depletion of HP1γ, Muchardt found that approx. 10% of the genes were regulated by HP1γ at the level of splicing. In a fraction of these genes, the H3K9me3 mark correlated with HP1γ-dependent splicing. Muchardt proposed a model in which HP1γ functions as a bridge between the chromatin and the pre-mRNA, mediating the transient contact between these two (a concept he called “Velcro-matin”). In his model, phosphorylated HP1γ can read the methylation marks of the CD44 region and creates contacts with the recently synthesized pre-mRNA, generating some kind of structure that slows down the RNA polymerase, increasing its accumulation at that region. The slowing down of the polymerase would in turn facilitate the recruitment of the spliceosome and the inclusion of the exon variant.

When the attention moved to the structure and organization of the nucleus, Peter Fraser, from The Babraham Institute in Cambridge, UK, and David Spector, from Cold Spring Harbor Laboratory, NY, discussed their work concerning the spatial organization of co-regulated genes and the nuclear organization of the H3K27 demethylase JMJD3, respectively. Fraser talked about RNA polymerase II transcription factories. He explained that active gene transcription is not continuous but occurs in bursts or pulses. Moreover, 99% of the genes appear to spend more time “off” than “on.” In cells taken directly from mice, transcriptional active genes share 200–400 RNA polymerase II-enriched factories. These shared nuclear foci have been postulated to contribute to increased local concentrations of specific factors required for gene expression, being optimized sites for the transcription of specific genes. Fraser showed results from genome-wide studies from his lab indicating that mouse α- and β-globin genes (Hba and Hbb) associated with hundreds of other genes that were regulated by the erythroid-specific Kruppel-like transcription factor, Klf1. Klf1 is mainly cytoplasmic but, remarkably, the few nuclear Klf1 foci overlapped with RNA polymerase II factories. Klf1 factories represented only 20% of the total number of factories but contain approximately 70% of the β-globin signal. Importantly, knock out of Klf1 disrupted nuclear organization of Klf1-dependent genes. In these cells, reintroduction of Klf1 led to rapid induction of Klf1-RNA polymerase II factories. Fraser estimated that at a given moment each factory is associated with 10–30 genes. Spector talked about the nuclear organization and dynamics of the H3K27 demethylase JMJD3. JMJD3 belongs to the jumonji C family of proteins, demethylates H3K27me2 and H3K27me3, and was shown to function as a tumor suppressor. Spector and colleagues studied the nuclear distribution of JMJD3 and observed that it showed a punctate nuclear distribution consisting of 20–40 foci of 0.2–0.5 micrometers in diameter. The localization of JMJD3 to these nuclear foci was dependent on an intact N-terminus and Zn finger domain. By analyzing the composition of JMJD3 nuclear bodies, they determined that they do not coincide with any of the previously identified nuclear bodies, nor were they enriched in euchromatin histone H3 marks or elongating RNA polymerase II. They instead showed significant co-localization with H3K9me3 and with facultative heterochromatin (containing HP1γ) in interphase nuclei. JMJD3 foci were absent between prometaphase and early G1, when the protein was diffuse. Using an artificial inducible system in living cells, they observed that JMJD3 was rapidly and transiently recruited to the activated transcription locus upon activation. Spector concluded that JMJD3 appears to be involved with the critical period of time when a gene is being induced.

Participants of this year's Symposium on Nuclear Organization & Function enjoyed six days of outstanding presentations, stimulating conversations and informal chats over the always welcomed glass of wine in the idyllic seaside background of the Cold Spring Harbor Laboratory. Exciting things keep happening in the nucleus and, luckily, we got to hear some of them at this meeting.

note

Speakers mentioned in this report approved discussion of their data.