Chromosome Research (2008) 16:397–412 # Springer 2008 DOI: 10.1007/s10577-008-1237-3 Elucidating chromatin and nuclear domain architecture with electron spectroscopic imaging David P. Bazett-Jones1*,RenLi1, Eden Fussner1, Rosa Nisman1 & Hesam Dehghani2 1Program in Genetics and Genome Biology, The Hospital for Sick Children, Research Institute, 101 College Street, East Tower, 15th Floor, 15-401T, Toronto, ON M5G 1L7, Canada; Tel: +1-416-813-2181; E-mail: [email protected]; 2Department of Physiology, School of Veterinary Medicine, Ferdowsi University of Mashhad, Mashhad, Iran The authors dedicate this paper to the memory of Ying Ren (1961Y2007). We all benefited from knowing her. Our research advanced through the technical creativity she provided. *Correspondence Key words: chromatin, correlative microscopy, electron microscopy, heterochromatin, nuclear speckles, nuclear structure, nucleolus, promyelocytic leukemia, transcription Abstract Electron microscopy has been the Fgold standard_ of spatial resolution for studying the structure of the cell nucleus. Electron spectroscopic imaging (ESI) offers advantages over conventional transmission electron microscopy by eliminating the need for heavy-atom contrast agents. ESI also provides mass-dependent and element-specific information at high resolution, permitting the distinguishing of structures that are primarily composed of protein, DNA, or RNA. The technique can be applied to understand the structural consequences of epigenetic modifications, such as modified histones, on chromatin fiber morphology. ESI can also be applied to elucidate the multifunctional behavior of subnuclear Forganelles_ such as the nucleolus and promyelocytic leukemia nuclear bodies. Abbreviations SR serine-arginine SUMOylation small ubiquitin-like modifier post-translational CBP Creb-binding protein modification CC condensed chromatin CTEM conventional transmission electron microscopy Introduction Daxx death-associated protein 6 ES murine embryonic stem cell ESI electron spectroscopic imaging The cell nucleus is composed of several compart- GFP green fluorescent protein ments including chromosome territories, nucleoli, ICS interchromosome space nuclear speckles (interchromatin granule clusters), IGC interchromatin granule clusters replication and transcription factories, promyelocytic ING1 inhibitor of growth family member 1 leukemia nuclear bodies (PML NBs), and a number MEL murine erythroleukemia cell Mdm2 murine double minute protein 2 of other bodies and assemblies. All await further PML NB promyelocytic leukemia nuclear body analysis and functional characterization. How these Rb retinoblastoma structures within the nucleus are organized appears to 398 D. P. Bazett-Jones et al. be largely a function of the various nuclear activities, energy losses that arise from core loss ionizations are such as transcription or replication of DNA (Kosak & discrete and are element-specific. The incident elec- Groudine 2004), yet the fundamental rules governing trons, which cause these events, could therefore be the relationship of nuclear structure to function used to generate element-specific images if it were remain to be elucidated. possible to filter or select these electrons from the Nuclear compartments are assembled and main- electron energy loss spectrum. Electron imaging spec- tained in a dynamic fashion (Misteli 2001), likely trometers have been designed and implemented for dictated by their function and involvement in nuclear this purpose. These devices accomplish two functions. processes. Nucleolar organization, for example, pro- First, in functioning as an electron spectrometer, they vides a unique spatial clustering of ribosomal genes separate electrons according to their energy losses. from different chromosomes to a defined nuclear This creates an electron energy loss spectrum. Their volume that enriches all of the players involved in second function is to remove aberrations created by transcription of ribosomal RNA genes. In another the spectrum-forming operation, and to create an example, accumulation of a set of active or poten- aberration-corrected image of the specimen. The two tially active genes likely contributes to the assembly functions together create an aberration-corrected of Ftranscription factories_, structures that provide a element-specific image. local permissive microenvironment for regulation Electron energy loss imaging, which we call and activity of RNA polymerase II (Osborne et al. electron spectroscopic imaging (ESI) offers advan- 2004). Our understanding of the relationships of tages over CTEM. First, the removal of electrons that chromatin, gene activation and subnuclear domains do not interact with the specimen as well as the has progressed with technical developments in cell selection of electrons that have undergone a specific imaging. Conventional transmission electron micros- energy loss through specimen interactions produces a copy (CTEM), for example, has revealed dense dark-field image. Such images can provide a great packing of chromatin, thought to be heterochromatin, increase in contrast over a bright-field CTEM image. along the nuclear envelope and at the periphery of Whereas heavy-atom contrast agents, such as uranium the nucleolus. In parallel, advances in light micros- and lead salts, are required to provide sufficient copy over the past two decades have revealed other contrast of poorly scattering light elements, i.e. those structural features such as the territorial organization elements that are relevant to biological material, ESI of chromosomes and dynamic features of both nucle- does not require these contrast-enhancing stains. This ar bodies and specific chromosome loci. However, is important for two reasons. The contrast agents the temptation to infer molecular detail from typical function by coating the biological material with large light microscopy observations, limited to 200 nm, precipitates of stain, consequently limiting the spatial can lead to over-interpretation of data, which in turn resolution. Second, these heavy atom salts do not leads to erroneous conclusions about nuclear structure interact uniformly with each biochemical component; and function (Dehghani et al. 2005). Efforts are now some compounds bind large amounts of uranium underway to fill this resolution gap between the con- acetate, for example, whereas other biochemicals do ventional light microscope and the electron microscope. not interact at all with the salt. This non-uniformity can create false impressions of Felectron-dense_ regions, and mass-depleted regions. In fact, the true Electron spectroscopic imaging mass density may not be reflected accurately by the uranium salt interactions. Electron spectroscopic imaging (ESI) has helped us A second advantage of ESI is that element-specific to address several diverse questions in both molec- maps can be created. This opportunity can be ular and cellular biology (Bazett-Jones & Hendzel exploited in cell-free studies of purified ribonucleo- 1999, Bazett-Jones et al. 1999, Dellaire & Bazett- protein (RNP) structures or DNA:protein complexes. Jones 2004). The technique is based on the principle The nucleic acid component contains significantly that some incident electrons that pass through a more phosphorus than the protein component of such specimen in the transmission electron microscope complexes. Phosphorus maps, therefore, provide a will lose energy due to ionization of both valence means of distinguishing the nucleic acid distribution shell and core electrons of the specimen_s atoms. The in such complexes. Moreover, quantitative evaluation Electron spectroscopic imaging of the nucleus 399 of these phosphorus-specific maps provides a mea- embryo (Figure 1). To prepare the sample for electron sure for nucleic acid in the complex. Mass-sensitive microscopy, the embryo was removed from the mother images, recorded in regions of the energy loss and fixed for 30 min in 4% paraformaldehyde. For spectrum that do not correspond to phosphorus, immunofluorescence or pre-embedding immunogold provide total mass information. Comparisons of the labeling, samples are treated with 0.5% Triton X-100. phosphorus maps and the mass images can therefore The detergent will extract membranes. The sample is be used to calculate stoichiometric relationships then post-fixed with 2% glutaraldehyde before resin between the nucleic acid and protein components of embedding. the complex. For these reasons, ESI has been applied The contrast in the low magnification image to the study of a variety of DNA: protein structures (Figure 1A) is proportional to the mass density of that are relevant to gene regulation. For example, the specimen. At this magnification and position in important architectural features of transcription fac- the energy loss spectrum (155 eV) only the con- tor interactions with promoter elements have been densed chromatin at the nuclear envelope and the observed (Bazett-Jones et al. 1994, Brown et al. well-contrasted nucleoli are clearly discerned. By 1996, Stefanovsky et al. 2001). At the level of recording images before and after the phosphorus chromatin organization, the importance of specific ionization edge (LII,III, 132 eV) (Bazett-Jones & chromatin modifications that destabilize nucleosome Hendzel 1999), a net phosphorus image is obtained structure (Bazett-Jones & Hendzel 1999), and how a through dividing the post-edge image, recorded at nucleosome remodeling complex can provide access 155 eV, by a pre-edge image, recorded at 120 eV. In to DNA in the nucleosome