Beyond the Sequence: Cellular Organization of Genome FunctionIntroductionCellular Organization of Nuclear ProcessesThe Organization of TranscriptionThe Dynamic Nature of Transcription ComplexesOrganization of DNA-Replication and -Repair SitesThe Stochastic, Self-Organizing Nature of Nuclear ProcessesHigher-Order Chromatin OrganizationChromatin as an Accessibility BarrierGenome Regulation via Local Chromatin LoopsThe Emergence of Large-Scale Chromatin LoopsSpatial Organization of GenomesInternal versus Peripheral Genome PositioningRelative Positioning: The Power of ProximityModels of Cellular Organization of Genome FunctionDeterministic ModelsSelf-Organization ModelsConclusionsAcknowledgmentsReferencesLeading EdgeReviewBeyond the Sequence: Cellular Organizationof Genome FunctionTom Misteli1,*1National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA*Correspondence: [email protected] 10.1016/j.cell.2007.01.028Genomes are more than linear sequences. In vivo they exist as elaborate physical struc-tures, and their functional properties are strongly determined by their cellular organization.I discuss here the functional relevance of spatial and temporal genome organization at threehierarchical levels: the organization of nuclear processes, the higher-order organization ofthe chromatin fiber, and the spatial arrangement of genomes within the cell nucleus. Recentinsights into the cell biology of genomes have overturned long-held dogmas and have led tonew models for many essential cellular processes, including gene expression and genomestability.IntroductionWe usually think of genomes abstractly as one-dimen-sional entities that are purely defined by their linear DNAsequences. Reality, of course, is far more complex. TheDNA helix is folded hierarchically into several layers ofhigher-order structures that eventually form a chromo-some (Woodcock, 2006). In this way, DNA is compactedand can be accommodated in the limiting space of thecell nucleus. The spatial arrangement of the chromatin fi-ber and the genome as a whole dramatically affects thefunction of DNA, and knowing the sequence of a genomeis insufficient to understand its physiological function.In addition to the complex arrangement of the geneticinformation itself, the cellular factors that read, copy, andmaintain the genome are organized in sophisticated pat-terns within the cell nucleus (Lamond and Spector, 2003;Misteli, 2005). Many transcription factors, chromatin pro-teins, and RNA-processing factors are compartmentalizedand accumulate in distinct nuclear domains; specificnuclear processes such as transcription and replicationoccur at spatially defined locations in the nucleus. Theorganizational properties of genomes and the machineriesthat act on them create an elaborate architectural environ-ment in which genomes must function. How they do so isone of the great challenges in modern cell biology.Uncovering the cell biology of genomes is fundamental.Although comparative genome analysis and large-scalemapping of genome features have yielded insights intothe physiological role of genetic information, these effortsshed little light onto the Holy Grail of genome biology,namely the question of how genomes actually workin vivo. The elucidation of the cellular organization of ge-nomes and its impact on genome regulation is a logicalnext step after the completion of sequencing projects.Understanding genome function within its architecturalframework is also highly relevant for biotechnologicalapplications that range from stem cell differentiation tosomatic cloning and gene therapy as all of these processesinvolve massive reorganization of nuclear architecture.Knowledge of the functional interplay between genomeorganization and activity will significantly contribute tomaking these applications more efficient and controllable.Cellular organization of genome function occurs at threehierarchical levels: the spatial and temporal organizationof nuclear processes themselves, including transcription,RNA processing, DNA replication, and DNA repair; theorganization of chromatin into higher-order domains;and the spatial arrangement of chromosomes and geneswithin the nuclear space. Each one of these levels hasregulatory potential, and all are interdependent. Severalsimple questions serve as guideposts to unravel the com-plex structure-function interplay of the genome in the cell:How are genome processes and genomes organized in3D space? What are the fundamental principles of organi-zation? What are the molecular mechanisms that give riseto the organization patterns? What are the physiologicalconsequences of spatial genome organization? Emerginganswers to these questions are now leading to unprece-dented insights into genome biology and to new, un-expected models of genome function.Cellular Organization of Nuclear ProcessesA hallmark of many nuclear processes is their spatial com-partmentalization. Most nuclear events do not occur ubiqui-tously throughout the nucleus but are limited to specific,spatially defined sites that often occur in dedicated nuclearbodies (Lamond and Spector, 2003; Misteli, 2005). Remark-ably, common mechanisms appear to organize some of thevastly different, fundamental nuclear processes.The Organization of TranscriptionThe most fundamental of all genome functions is tran-scription. Surprisingly, there is still much uncertainty asto how transcription is organized within the nucleus(Cook, 1999; Chakalova et al., 2005). Visualization ofCell 128, 787–800, February 23, 2007 ª2007 Elsevier Inc. 787transcription sites reveals the presence of several thou-sand distinct sites that appear to be randomly dispersedthroughout the nuclear volume (Wansink et al., 1993;Figure 1A). Influenced largely by in vitro analysis of thetranscription machinery, it was long assumed that this dis-tribution represents RNA polymerases (RNA pol) elongat-ing along genes. But an alternative and increasingly plau-sible view is that these sites correspond to subnucleartranscription centers (Cook, 1999; Chakalova et al.,2005; Figure 1A). As originally proposed by Cook, these‘‘transcription factories’’ are transcription hot spots thatharbor enough transcription factors and polymerases toserve multiple genes (Cook, 1999). The organization oftranscription in centralized structures that contain multipletranscription machineries is consistent with the presenceof an estimated 65,000 active RNA pol II molecules butfewer than