Transcription factor binding to a DNA zip code controls interchromosomal clustering at the nuclear periphery

Abstract

Active genes in yeast can be targeted to the nuclear periphery through interaction of cis-acting "DNA zip codes" with the nuclear pore complex. We find that genes with identical zip codes cluster together. This clustering was specific; pairs of genes that were targeted to the nuclear periphery by different zip codes did not cluster together. Insertion of two different zip codes (GRS I or GRS III) at an ectopic site induced clustering with endogenous genes that have that zip code. Targeting to the nuclear periphery and interaction with the nuclear pore is a prerequisite for gene clustering, but clustering can be maintained in the nucleoplasm. Finally, we find that the Put3 transcription factor recognizes the GRS I zip code to mediate both targeting to the NPC and interchromosomal clustering. These results suggest that zip-code-mediated clustering of genes at the nuclear periphery influences the three-dimensional arrangement of the yeast genome.

Figure 1. INO1 is targeted to a limited region of the nuclear envelope

(A) Representative micrograph of yeast nuclei having the lac repressor array integrated at INO1 and expressing GFP-LacI, GFP-Nup49 labeling the nuclear envelope and Nop1-mCherry labeling the nucleolus (maximum projection of 250nm Z stacks). Scale bar = 1µm. (B) Coordinates used in this study: nucleolus and nucleus centroids (red and grey balls respectively), central axis (dashed line joining the two centroids), radial axis (dashed line joining the locus and the nuclear centroid), α, polar angle defined by these two axes. (C) INO1 gene map based on the analysis of 2950 nuclei grown in the presence of inositol 100µM (repressed, left) or on the analysis of 2627 nuclei grown in the absence of inositol (active, right). Dashed red circle: ‘median’ nucleolus; red circle, median location of nucleolar centroid. The color scale indicates the probability that the locus is inside the region enclosed by the corresponding contour, which may include regions enclosed by other contours. The dark red contour corresponds to the localization observed in 10% of the population and the dark blue contour corresponds to the localization observed in 90% of the population. See also Figure S1.

Figure 2. INO1 clustering

(A) Schematic of experimental strategy. An array of 128 Lac repressor-binding sites was integrated beside URA3 on chromosome V and an array of 112 Tet repressor binding sites was integrated beside INO1 on chromosome X in a strain expressing GFP-LacI and RFP-TetR. The positions of GRS I and GRS II in the promoter of INO1 are indicated as I and II, respectively. To create URA3:INO1, the INO1 gene was integrated beside URA3. (B) Example distances between two loci. Cells were fixed, stained with antibodies against GFP and RFP, visualized by line-scanning confocal microscopy and distances between the centers of the spots were measured using Zeiss LSM software. Scale bar = 1µm. For panels C–G: Left: plot of the distribution of the frequency of each distance between loci in the population (n ≥100 cells). Right: the fraction of cells in which the loci were < 0.55µm apart. (C) Distribution of distances between active INO1 and URA3 or between active INO1 and URA3:INO1. (D) Distribution of distances between INO1 and URA3:INO1 under activating (−inositol; same distribution as in panel C) and repressing (+inositol) conditions. (E) Distribution of distances between either two alleles of URA3 or two alleles of INO1 in diploid cells in repressing conditions. (F) Distribution of distances between two alleles of INO1 in diploid cells in repressing or activating conditions. (G) Distribution of distances between INO1 and GAL1 in cells grown in activating conditions for INO1 and either repressing (glucose) or activating (galactose) conditions for GAL1.

Figure 3. Gene clustering is specific

(A) Left: schematic of two green dot experimental strategy. An array of 256 Lac repressor-binding sites was integrated beside HSP104 and an array of 128 Lac repressor-binding sites was integrated beside other loci. Right: representative confocal micrographs of a strain having a large array and a small array, stained with anti-GFP and anti-myc (to stain myc-tagged Sec63 in the endoplasmic reticulum and nuclear envelope). Scale bar = 1µm. For panels B, C & F, Top: chromosomal locations of genes. Left: plot of the distribution of the distances between loci in the population. Right: the fraction of cells in which the loci were <0.55µm apart. (B) Distribution of distances between HSP104 and INO1, under conditions in which only INO1 is active (−inositol) or conditions in which both genes are active (−inositol + 10% ethanol). (C) Distribution of distances between HSP104 and GAL1, under uninducing (glucose) or inducing (galactose +10% ethanol) conditions. (D) Representative images scored for co-localization of HSP104 or GAL2 with either the nuclear envelope (stained with anti-myc for Sec63-myc) or the nucleolus (stained with anti-Nop5/6). Scale bar = 1µm. (E) The fraction of cells in which HSP104 or GAL2 co-localized with the nuclear envelope or the nucleolus (n = 3; 30–50 cells per biological replicate; error bars = SEM). (F) Distribution of distances between HSP104 and GAL2, grown in uninducing or inducing conditions.

Figure 4. Clustering of INO1 requires targeting to the nuclear pore complex

(A) Distribution of distances between INO1 and URA3:INO1 in NUP2 (data from Figure 2C) and nup2Δ cells grown under activating conditions (−inositol). (B) Clustering of active INO1 and URA3:INO1 in cells in which the nuclear envelope was also stained. Clustering of the two loci was determined in ~30 cells of each of the three classes: both genes on the nuclear envelope (on/on), one gene on the nuclear envelope (on/off) or both genes off the nuclear envelope (off/off). In all cases where the locus was scored as peripheral, the center of the green dot was < 0.25µm from the cytoplasmic edge of the nuclear envelope. (C) Cells having the Tet repressor array at INO1 and the Lac repressor array at URA3:INO1 were arrested with hydroxyurea either after activating INO1 or before activating INO1.

Figure 5. Gene clustering is controlled by DNA zip codes

Panels A & B used the strategy described in Figure 2A. Panels C & D used the strategy described in Figure 3A. (A) Distribution of distances between URA3:INO1 and either wild type INO1 (same data as Figure 2C) or grsImut INO1 (Ahmed et al., 2010) in cells grown under activating conditions (−inositol). (B) Distribution of distances between active INO1 and either URA3:INO1 (same as Figure 2C) or URA3:GRS I. (C) Distribution of distances between active HSP104 and either URA3 or URA3:HSP104prom. (D) Distribution of distances between HSP104 and URA3:GRS III in cells grown in the presence or absence of 10% ethanol. See also Figures S2 and S3.

Figure 6. Clustering of endogenous GRS I-targeted genes

(A) Distribution of distances between TSA2 and INO1 in cells grown under activating conditions for INO1 (−inositol) or activating conditions for INO1 and TSA2 (−inositol +10% ethanol). (B) Distribution of distances between TSA2 and either wild type INO1 (same data as in panel A) or grsImutINO1 in cells grown under activating conditions.

Figure 7. The Put3 transcription factor mediates GRS I-dependent clustering

(A) An electrophoretic mobility shift assay of yeast lysates incubated with radiolabeled 4×GRS probe. Lysates were prepared from either a wild type strain (lanes 1 & 2) or a put3Δ mutant strain (lanes 3–6). The put3Δ strain was transformed with a plasmid expressing GST-PUT3 under the control of the ADH1 promoter (lanes 5&6). Anti-GST antibody was added to reactions in lanes 2, 4 and 6. (B) Localization of repressed and active INO1 and URA3:INO1 in PUT3 and put3Δ mutants with respect to the nuclear envelope. The dynamic range of this assay is from 20%–85% and the blue, hatched line represents the distribution of the URA3 gene with respect to the nuclear envelope (Brickner et al., 2010; Brickner and Walter, 2004). (C) Localization of TSA2 in PUT3 and put3Δ cells before or after heat shock (30 minutes). (D) ChIP against GST or GST-Put3 from cells grown in the presence or absence of inositol. (E) ChIP against Nup2-TAP from PUT3 and put3Δ strains grown in the presence or absence of inositol. For panels C – E, the immunoprecipitated DNA was quantified relative to input by real-time quantitative PCR. (F) mRNA levels for INO1, URA3:INO1 or URA3:grsImut INO1 in PUT3 or put3Δ strains, quantified by RT-qPCR relative to ACT1 mRNA. For panels B–F, error bars = SEM. (G) Left: distribution of distances between Tet repressor-marked INO1 and Lac repressor-marked URA3:INO1 in wild type (same data as in Figure 2A) and put3Δ cells grown under activating conditions (−inositol). Right: the fraction of cells in which the loci were < 0.55µm apart. See also Figure S4.