Genome organization around nuclear speckles

Abstract

Higher eukaryotic cell nuclei are highly compartmentalized into bodies and structural assemblies of specialized functions. Nuclear speckles/IGCs are one of the most prominent nuclear bodies, yet their functional significance remains largely unknown. Recent advances in sequence-based mapping of nuclear genome organization now provide genome-wide analysis of chromosome organization relative to nuclear speckles. Here we review older microscopy-based studies on a small number of genes with the new genomic mapping data suggesting a significant fraction of the genome is almost deterministically positioned near nuclear speckles. Both microscopy and genomic-based approaches support the concept of the nuclear speckle periphery as a major active chromosomal compartment which may play an important role in fine-tuning gene regulation.

1.. Variable Identification/Visualization of Nuclear Speckles.

(a) Cajal identified nuclear speckles (handwritten “a”) using different histochemical stains, including silver staining; also shown are nucleoli (handwritten “b”). Original drawing at the Cajal Institute, CSIC, Madrid. Reprinted from Fig. 2a, Ref. [3]. (b) Energy loss EM visualization of nuclear speckles in Hela cells as Interchromatin Granule Clusters (green arrows) surrounded by chromatin (red arrows); energy loss signal (black) is proportional to presence of nucleic acid (phosphorus). Image courtesy of Michael Hendzel. (c) EM section showing nuclear speckles in detergent-extracted nucleus stained with heavy metals in CHO cells; a peripheral zone (large arrowheads), which appears distinct from the surrounding chromatin (narrow arrows), surrounds the granular core of the speckle. Negative image is shown, with heavy metal staining appearing white and low staining black. Scale bar= 0.5 μm. Reprinted from Fig. 5B, Ref. [8]. (d) lncRNA Malat1 (red) and U2 snRNA (green) concentrate at the speckle periphery surrounding SC35 (blue) speckle core as visualized by 3D SIM light microscopy; U2 snRNA also localizes to small foci outside of speckles. Diffraction-limited image (top left) is shown for comparison. Reprinted from Fig. 1A, Ref. [9]. (e) The SON speckle marker (left) shows a higher, more localized concentration over nuclear speckles than SR splicing factors such as ASF/SF2 (right) which stains a larger region surrounding speckles, regions connecting speckles, and other focal accumulations. Scale bar= 2 μm. Reprinted from Fig. S1A, Ref. [24].

2.. Poly(A)+ RNA, Decondensed Chromatin, and Transcription Sites at the Nuclear Speckle Periphery.

(a) Nuclear speckles were identified as “Poly(A)+ RNA islands”(red) and “SC35 domains” (green); the Poly(A)+ RNA signal overlaps but is also peripheral to the SC35 domain. Reprinted from Fig. 2A,E,F, Ref. [2]. (b) Serial-section EM reconstruction of detergent-extracted nucleus stained with heavy metals in CHO cells. Negative image is shown, with heavy metal staining appearing white and low staining black. Locally decondensed large-scale chromatin fibers (small arrowheads) frequently mapped to the Interchromatin Granule Cluster periphery (large arrowheads). Scale bar= 120 nm Reprinted from Fig. 8E, Ref. [8]. (c-e) Active sites of transcription and Ser5p- RNA pol2 staining foci surround nuclear speckle periphery. (d-e) 5 min EU pulse-labeling of transcription (d) or Ser5p-RNA pol 2 staining (e) (red) relative to nuclear speckle marker (SON) (green in (c) and white in (d)). 2x enlarged views corresponding to regions surrounding speckles (arrows). Reprinted and Modified from Fig. 5C&D, Ref. [24, with permission from Rockefeller University Press.

3.. Genomic assays for speckle localization and correlation with speckle localization measured by FISH.

(a) MARGI uses an oligonucleotide adaptor to ligate RNA and DNA fragments (top left). Counts of ligations with nuclear-speckle associated RNAs (saRNAs) show peaks and valleys across chromosomes (bottom left). FISH measurements show an increased level of SC35 colocalization with probes selected from peaks (nsaPeak) versus valleys (non-nsaPeak). Adapted under Creative Commons Attribution License (CC BY) from Graphical Summary, Fig. 4A,D, Ref. [37], (b) SPRITE measures colocalization of DNA and RNA fragments in isolated, chromatin complexes (top left). Large complexes with many DNA fragments show unusually high numbers of inter-chromosome interactions with a small number of active chromosomal regions, defining an “active chromosomal hub”. The frequency of chromosome interactions with this active hub (red, y-axis) correlate with Malat1 RAP-RNA-Seq (black, y-axis) (bottom left). FISH of different probes showed a linear correlation in percentage of alleles mapping <0.5 μm from a nuclear speckle (y-axis) with this active hub contact frequency (x-axis) (right). Adapted from Graphical Summary, Figs. 5A, 6D, Ref.[36]. (c) TSA-Seq measures spreading of diffusible tyramide free-radical, generated by a peroxidase localized near nuclear speckles by SON immunostaining, producing biotin-DNA labeling proportional to speckle distance. DNA purification is followed by biotinylated-DNA pull-down and sequencing (top left). TSA-Seq score uses log2 ratio of observed versus average read count (y-axis, log2 ratio of fold-enrichment), with peaks predicting regions near speckles (bottom left). An exponential relationship was observed between probe TSA-Seq fold-enrichment and the average FISH-signal distance from a speckle (right). Adapted from Figs. 1A, 2A,B, 3A,B, Ref. [24] with permission from Rockefeller University Press.

4.. Models for explaining variable speckle proximity/distance observed by genomic methods.

Variable size peaks observed for SPRITE [36], TSA-Seq [24], and MARGI [37] have been interpreted differently. (a) Model 1: SON TSA-Seq identified two types of transcription hot-zones mapping near TSA-Seq local maxima: Type I (large peaks, correlated with A1 Hi-C subcompartment) positioned close, on average, to nuclear speckles and Type II peaks (smaller peaks, correlated with A2 Hi-C subcompartment), positioned at intermediate distances, on average, from speckles, and possibly interacting with an unknown nuclear body. (b) Model 2: Larger TSA-Seq (Type I), MARGI, and SPRITE peaks could correspond to a larger fraction of alleles localizing close to nuclear speckles, with smaller TSA-Seq (Type II), MARGI and SPRITE peaks showing a smaller fraction of alleles close to speckles. (c) Model 3: Larger (Type I) versus smaller (Type II) TSA-Seq peaks could correspond to association of chromosome region with larger versus smaller nuclear speckles, respectively. Further work is needed to determine which model best describes chromosome positioning relative to nuclear speckles.