Nuclear GPS for interchromosomal clustering

Summary

Nuclear architecture and the relative position of a gene can play roles in the regulation of its expression. Brickner et al. (2012) now analyze nuclear global positioning of genes and reveal that the Put3 transcription factor functions with cis-encoded DNA elements and nuclear pore complexes to regulate interchromosomal gene clustering.

As a cell differentiates and undergoes distinguishing gross morphological changes, much more is happening than meets the eye: The organization of DNA also changes dramatically. Functionally, nuclear “global” positioning of a particular gene in a cell lineage during development is thought to reflect whether the gene is primed for activation or repression (Schoenfelder et al., 2010; Meister et al. 2010). Classic studies revealing that chromosomes were confined to select nuclear regions (Zorn et al., 1979) have fueled an interest in determining whether the information for such precise arrangements is encoded within the DNA itself. Although it is now clear that there are additional layers of regulatory complexity that influence gene positioning, the elaborate mechanisms for establishing such higher order chromatin organization are only beginning to be elucidated. One frequently observed nuclear arrangement, termed interchromosomal clustering, suggests that the spatial positioning of genes together reflects a shared mode of transcriptional control (Schoenfelder et al., 2010; Xu and Cook, 2008). Excitingly, Brickner et al. (2012) report in this issue of Developmental Cell the analysis of both DNA and protein determinants that modulate clustering of genes from different chromosomes to the same region at the nuclear periphery of the budding yeast Saccharomyces cerevisiae.

The S. cerevisiae nucleus is a robust model for studying nuclear organization. It has three distinct subdomains – the nucleoplasm, the nuclear periphery, and the nucleolus (Figure 1A). The nuclear periphery can be further broken down into distinct repressive and active zones for gene expression. While centromere and telomere anchoring sites on the periphery generally contact silenced regions of the genome, the intervening regions are occupied by nuclear pore complexes (NPCs) and provide a permissive environment for active gene expression (Figure 1A). There are several examples of genes moving from the nucleoplasm, where they are inactive, to the periphery for transcriptional activation in response to changes in nutrient availability and temperature, and for programmed cell morphological changes during the yeast-mating pathway (reviewed in Zimmer and Fabre, 2011).

Figure 1. Interchromosomal clustering to subnuclear regions in S. cerevisiae and metazoans.

A. In S. cerevisiae, the nucleus is composed of a nuclear periphery (red), a nucleolus (red outlined in dashed line) and the nucleoplasm (pink). The nuclear pore complex (NPC) and surrounding local environment are predicted to provide a subdomain permissive for gene expression (light green outlined in dashed line). Changes in the environment trigger genes with similar GRS elements to cluster together at the nuclear periphery (dark green outlined in dashed line). The localization mechanism requires the transcription factor Put3 and the nuclear pore complex (NPC) component Nup2. B. Metazoan nuclei have multiple nuclear bodies with distinct functions. The nuclear periphery (red) contains a heterochromatin and nuclear lamina meshwork alternating with heterochromatic exclusion zones and NPCs (light green outlined in dashed line). Nuclei can have from 1-4 nucleoli (red outlined in dashed line). In metazoans, nuclear rearrangements occur in response to environmental and developmental cues, wherein genes colocalize to specialized transcription factories (dark green outlined in dashed line). The molecular determinants for positioning in S. cerevisiae or metazoans remain to be fully elucidated (? symbol).

The Brickner group previously identified S. cerevisiae DNA ‘zip codes’ that are both necessary and sufficient for gene positioning at the nuclear periphery (Light et al., 2010). These zip codes are cis-encoded DNA elements found within the promoter regions of inducible genes. The INO1 gene harbors two well-defined zip codes, termed gene recruitment sequences (GRSI and GRSII), which position the gene at the nuclear periphery under activating conditions of inositol starvation. Interactions with specific components of the NPC are also necessary for GRS-mediated peripheral localization and optimal INO1 transcriptional induction (Light et al., 2010).

To further investigate the mechanism for INO1 gene positioning, Brickner et al. (2012) began by using clever genetic tools and comparisons to other GRS-containing genes. Strikingly, INO1 and TSA2 both contain GRSI and cluster to an overlapping region in the nuclear periphery. However, INO1 and HSP104, which contain distinct GRSI and GRSIII elements, do not. Overall, they show that shared GRS sequences are both necessary and sufficient to mediate specific interchromosomal clustering. Brickner et al. also find that peripheral gene targeting via interaction with the NPC is a critical step prior to interchromosomal clustering.

To tackle the more difficult task of identifying trans-acting protein determinants of gene localization, Brickner et al. (2012) speculated that if a protein has binding affinity for the GRSI sequence, then it could directly contribute to peripheral targeting and interchromosomal clustering. Through a DNA affinity purification scheme accompanied by mass spectrometry and genetic analysis, Brickner et al. honed in on Put3, a member of the Zn₂-Cys₆ zinc finger transcription factor family. They showed that it is necessary for GRSI-mediated peripheral targeting and interchromosomal clustering. Put3 is also required for NPC-interactions and optimal expression of the INO1 gene. Interestingly, Put3 regulates gene transcription via an unrelated UAS_PUT element (Siddiqui and Brandriss, 1989), suggesting that it has dual functions at promoters. Overall, these are important steps in defining the precise machinery at work in controlling gene localization.

Can these principles of subnuclear organization in S. cerevisiae be applied to metazoans? Although caution is needed given the unique aspects of nuclear architecture in metazoans compared to yeast (Figure 1B) (reviewed in Mao et al., 2011 and Zimmer and Fabre, 2011), unraveling the machinery for peripheral gene positioning in S. cerevisiae could give insight into how genes are localized to active sites of transcription (i.e. transcription factories) in metazoan nuclei (Shoenfelder et al., 2010; Xu and Cook, 2008). Indeed, the requirement for a specialized transcription factor in S. cerevisiae DNA zip code-mediated positioning corresponds directly with studies of the transcription factor Klf1 in erythroid cells (Schoenfelder et al., 2010). Similar to Put3, Klf1 influences the nuclear localization of coregulated genes to distinct transcription factories. Thus, one might predict that metazoan transcription factories are not random assemblies of active genes, but rather result from unique associations of genes with common DNA zip codes. In response to developmental and environmental cues, DNA zip codes might help define a metazoan subnuclear organization to support a specific transcriptional program. Interestingly, during development in C. elegans, tissue-specific promoters localize to the nucleoplasm coincident with transcriptional activation (Meister et al., 2010). Evidence in metazoans also points to key roles for NPC proteins in transcriptional regulation during differentiation (D’Angelo et al., 2012 and references cited therein).

The combination of a less complex genome and robust environmentally-controlled gene expression pathways in S. cerevisiae present an excellent system in which to pair single-cell based microscopy approaches with population-based genome-wide association analysis (as in metazoans, see Schoenfelder et al., 2010). This will achieve greater resolution of the functional regulation for these events and further elucidate the underlying machinery. For example, do different environmental responses require distinct interchromosomal clustering events with each involving a specific transcription factor? It is also unclear whether gene localization to nuclear subdomains is established through active localization machinery or through a passive mechanism of retention. Do NPCs provide a local environment that is permissive for gene expression or a coordinate in the nuclear global positioning system (GPS) for interchromosomal clustering? If so, Put3 might be part of a tethering scaffold between NPC components and GRSI-containing genes. Given that gene localization occurs in response to environmental cues, it is also tempting to speculate that Put3’s dual functions are modulated in a signaling-dependent manner. Ongoing studies will further discern how the non-randomness of gene clustering is linked to functional specificity. Ultimately, testing these principles in additional developmental systems and disease models will expand our understandings of context-specific genome architectures.