RNA polymerase I-specific subunits promote polymerase clustering to enhance the rRNA gene transcription cycle

Abstract

RNA polymerase I (Pol I) produces large ribosomal RNAs (rRNAs). In this study, we show that the Rpa49 and Rpa34 Pol I subunits, which do not have counterparts in Pol II and Pol III complexes, are functionally conserved using heterospecific complementation of the human and Schizosaccharomyces pombe orthologues in Saccharomyces cerevisiae. Deletion of RPA49 leads to the disappearance of nucleolar structure, but nucleolar assembly can be restored by decreasing ribosomal gene copy number from 190 to 25. Statistical analysis of Miller spreads in the absence of Rpa49 demonstrates a fourfold decrease in Pol I loading rate per gene and decreased contact between adjacent Pol I complexes. Therefore, the Rpa34 and Rpa49 Pol I-specific subunits are essential for nucleolar assembly and for the high polymerase loading rate associated with frequent contact between adjacent enzymes. Together our data suggest that localized rRNA production results in spatially constrained rRNA production, which is instrumental for nucleolar assembly.

Figure 1.

PAF53/CAST heterodimer is functionally conserved from human to budding yeast. (A) Pol I–specific subunits in S. cerevisiae, S. pombe, Mus musculus, and Homo sapiens. (B) Nucleolar localization of CAST depends on PAF53. CAST fused with YFP (YFP-CAST) is expressed in a WT yeast producing Nop1 fused to mCherry (mCherry-Nop1) with (PAF53) or without (−) PAF53 coexpression (see Materials and methods). (right) The overlay of both fluorescent signals (merge) is shown. (C) Immunoprecipitation of budding yeast Pol I demonstrates that association of CAST to Pol I requires human PAF53. HA-tagged CAST (HA-CAST) is detected in the supernatant (SUP) with or without coexpression of PAF53. Using IgG Sepharose to immunopurify TAP-tagged budding yeast Pol I (IP), CAST copurifies with Pol I only when coexpressed with PAF53 (+). Western blot using Pap and 12CA5 revealed tagged budding yeast Pol I (Rpa190-TAP) and CAST (HA-CAST) in supernatant (SUP) and immunoprecipitated fraction (IP), respectively. (D) Nucleolar localization of PAF53 depends on CAST. PAF53 fused with YFP (YFP-PAF53) is expressed in a WT budding yeast producing Nop1 fused to mCherry (mCherry-Nop1) with (CAST) or without CAST (−). CAST expression is required for both a detectable YFP signal and a nucleolar localization of PAF53. (E) PAF53 expressed in budding yeast is stabilized by CAST coexpression. PAF53 is detected in a whole cell extract from WT budding yeast expressing PAF53 (PAF53) alone or coexpressed with CAST (PAF53 CAST) using PAF53 antibody (α-PAF53). Nop1 (α-Nop1) is used as loading control. Total extract from HeLa loaded in lane 1 is used as control for PAF53 detection. White lines indicate that intervening lanes have been spliced out. (F) Overexpressed PAF53 complements rpa49Δ cold-sensitive growth defect. 10-fold serial dilutions of RPA49-deleted strain expressing budding yeast RPA49 (Rpa49), empty vector (−), PAF53 expressed from a strong promoter (P_PGK-PAF53), or an even stronger budding yeast promoter P_TEF1 (P_TEF1-PAF53) were spotted on 5-FOA–containing plates to check for complementation of the mutant cold sensitivity at 25°C (see Materials and methods for plasmid-shuffling assay). Plates without FOA (−) are used as control to confirm that the same number of cells were spotted. The strength of the complementation is evaluated by comparing plates with (FOA) or without FOA (−). (G) Expression of the PAF53/CAST heterodimer complements the RPA34 deletion phenotype. 10-fold serial dilutions of rpa34 deletion strains containing budding yeast RPA34 (Rpa34), empty vector (−), PAF53 driven from a moderate promoter (PAF53), CAST expressed from the budding yeast promoter P_TEF1 (CAST) or coexpressing PAF53 and CAST (PAF53 CAST) were spotted on 1 g/liter caffeine-containing plates to check for suppression of the rpa34Δ mutant hypersensitivity. Strength of the suppression of drug sensitivity was evaluated by comparing plates with (+) or without (−) caffeine spotted with equivalent amount of cells.

Figure 2.

Rpa49/Rpa34 heterodimer is conserved in fission yeast. (A) In S. cerevisiae, nucleolar localization of Sp-Rpa34 depends on Sp-Rpa49. Sp-Rpa34 fused with YFP (YFP-Sp-Rpa34) is expressed in a WT budding yeast with (Sp-Rpa49) or without (−) Sp-Rpa49 coexpression. In all cases, cells produce Nop1 fused to mCherry (mCherry-Nop1) and Nup49 fused with CFP (CFP-Nup49) to label the nucleolus and the nuclear pore complexes, respectively. (right) Overlay of the three fluorescent signals (merge) is shown. (B) Sp-Rpa34/Sp-Rpa49 heterodimer complements budding yeast Rpa34 essential function in a Pol I mutant background (rpa135-D395N). 10-fold serial dilutions of rpa34Δ rpa135-D395N double-deletion strains containing S. cerevisiae RPA34 (Rpa34), empty vector (−), Sp-Rpa34 (Sp-Rpa34), Sp-Rpa49 (Sp-Rpa49), or coexpressing Sp-Rpa34 and Sp-Rpa49 (Sp-Rpa34/Sp-Rpa49) were spotted on plates with (FOA) or without FOA (−), and complementation was evaluated by comparing plates with (FOA) or without FOA (−). (C) Sp-Rpa34 deletion barely affects growth in S. pombe. 10-fold serial dilutions of S. pombe WT and Sp-rpa34 deletion strains in rich medium at 30°C after 48 (left) and 120 h (right). (D) Sp-rpa34 Sp-rpa49 double deletion is viable in S. pombe. To generate the double-mutant strain, we crossed two haploid strains bearing the single Sp-rpa34 or Sp-rpa49 deletion. After meiosis, we analyzed the growth of single- and double-mutant offspring. Viability of the double mutant was confirmed in 20 tetrads. A representative tetratype tetrade is shown, with WT, single Sp-rpa34, or Sp-rpa49 deletion and double Sp-rpa34Δ Sp-rpa49Δ (ΔΔ) genotype. (E) The sensitivity of the Sp-rpa34 deletion strain to mycophenolic acid and 6-azauracil is suppressed by Sp-Rpa49 overexpression. 10-fold serial dilutions of Sp-rpa34Δ S. pombe expressing Sp-RPA34 (Sp-Rpa34), an empty vector (Sp-rpa34Δ), or overexpressing Sp-RPA49 (Sp-rpa34Δ + Sp-Rpa49) were spotted on rich media (left; −), 75 µg/ml 6-azauracil (middle; +6-AZA), or 40 µg/ml mycophenolic acid (right; +MPA) and grown for 4 d at 25°C.

Figure 3.

The C-terminal domain of Rpa34 is a nucleolar localization signal. (A) Rpa34 C-terminal domain is fully dispensable in vivo. 10-fold serial dilutions of four double-mutant budding yeast strains, rpa34Δ top1Δ (top left), rpa34Δ stb5Δ (top right), rpa34Δ rpa14Δ (bottom left), or rpa34Δ gcr2Δ (bottom right), are shown. Double mutants containing RPA34 (Rpa34), empty vector (−), Rpa34-ΔC driven from a P_pgk1 promoter (Rpa34-ΔC), or overexpressing Rpa49 (2µ-RPA49) were spotted on 5-FOA–containing plates to check for complementation of the lethality of the double mutant. Plates were incubated for 3 d at 30°C. Strength of the complementation was evaluated by comparing plates with (FOA) or without FOA (−). (B) Nucleolar localization of Rpa34 is independent of Rpa49. Rpa34 fused with YFP (YFP-Rpa34) is expressed in a WT budding yeast or in a mutant lacking Rpa49 (rpa49Δ). Both strains produce Nop1 fused to mCherry (mCherry-Nop1). (right) Overlay of both fluorescent signals (merge) is shown. (C) Nucleolar localization of Rpa34-ΔC is dependent on Rpa49. Rpa34ΔC fused with YFP (YFP-Rpa34ΔC) is expressed in WT or in a mutant lacking Rpa49 (rpa49Δ). Both strains produce Nop1 fused to mCherry (mCherry-Nop1). (right) Overlay of both fluorescent signals (merge) is shown.

Figure 4.

Rpa34 and Rpa49 are required for nucleolar integrity. (A) Nucleolar structure of rpa49Δ and rpa34Δ cells assessed by fluorescent microscopy. Representative confocal sections of strains TMS1-1a (WT), BEN18-1a (rpa34Δ), and BEN19-1a (rpa49Δ) expressing TetR-GFP staining the nucleus, GFP-Nup49 staining the NPC (NPC/nucleus), and mCherry-Nop1 staining the nucleolus (mCherry-Nop1) are shown. (right) Overlay of both fluorescent signals (merge) is shown. (B) Cumulative distribution function of the nucleolar volume in Pol I mutants. The nucleolar volume was estimated by thresholding using the intensity of the mCherry fluorescent signal in WT, rpa49Δ, and rpa34Δ mutant cells (n_WT = 907; n_rpa49 = 620; n_rpa34 = 928). (C) Ultrastructural study of cryofixed, cryosubstituted WT, rpa34Δ, and rpa49Δ budding yeast cells. Representative sections of the nuclei are depicted, with manual segmentation of nucleolus bounded by red lines. The cytoplasmic ribosomal contents in the three budding yeast strains are depicted in a selected frame. (D) Quantification of ultrastructural morphological alterations in Pol I mutants. Nucleolar area normalized by nuclear surface is estimated by manual segmentation of electron microscopy sections (n = 20). (E) Quantification of the ribosomal cytoplasmic content in Pol I mutants. Ribosomal content of representative sections (n = 20) was manually counted and expressed as ribosomes per micrometer cubed. (F) Polysomal profiles (OD₂₆₀ nm) after sucrose density gradient centrifugation derived from the rpa49Δ mutant and WT strain. The positions of 40S ribosomal subunits, 60S ribosomal subunits, polysomes, and half-mer polysomes (H) are indicated.

Figure 5.

rpa49Δ phenotype in strains with 190 or 25 copies of rDNA. (A) rDNA copy number is unaffected by rpa49 deletion in a fob1 deletion background. PFGE of chromosomes from rpa49Δ fob1Δ strains resulting from invalidation of the rpa49Δ gene in fob1Δ strains containing either 25 or 190 copies of rDNA or from original fob1Δ strains. Separated chromosomes were stained with ethidium bromide (right). Hybridization of rDNA probes was performed to estimate chromosome XII size (left). (B) Growth phenotype of the fob1Δrpa49Δ double-mutant strains bearing 190 or 25 copies of rDNA. 10-fold serial dilutions of strains BEN20-1a (25 rDNA copies rpa49Δ fob1Δ) and BEN21-1a (190 rDNA copies rpa49Δ fob1Δ) deleted for rpa49 and containing 25 or 190 copies, respectively, of rDNA and an empty vector (−) or bearing RPA49 (Rpa49) were spotted on plates with (FOA) or without FOA (−) and grown at 25°C for 7 d. (C) Nucleolar structure of Pol I mutants assessed by fluorescent microscopy. Representative confocal sections of strains NOY1071 (25 rDNA copies fob1Δ), NOY1064 (190 rDNA copies rpa49Δ), and BEN20-1a (25 rDNA copies rpa49Δ) containing GFP-Nup49 staining the NPC (GFP-Nup49) and mCherry-Nop1 staining the nucleolus (mCherry-Nop1). (right) Overlay of both fluorescent signals (merge) is shown. (D) Ultrastructural study of cryofixed, cryosubstituted strains with 25 rDNA copies with (WT) or without Rpa49 (Rpa49Δ).

Figure 6.

Quantification of polymerases along transcribed rDNA genes. (A) Visualization of active genes in rDNA. Using a mutant strain with a reduced number of rDNA copies (strain NOY886; 42 rDNA copies), we obtained high quality spreads from total nucleolar DNA. (B) Quantification of actively transcribed rDNA genes. The quantification of active genes in mutant strain (42 active gene in NOY886) is compatible with previous results (French et al., 2003). (C) Visualization of individual polymerases on negatively stained Miller spread. In mutant strains with 25 rDNA copies, Pol I density on each gene is very high. Using higher magnification (bottom left), we can still detect individual polymerases, which can be quantified (right; red circles). Note that a nascent rRNA can be visualized from each detected polymerase.

Figure 7.

Miller spread of 190 and 25 rDNA copy strains with or without Rpa49. (A) Representative Miller spread from NOY1071 bearing 25 rDNA copies (25 rDNA copies [WT]), BEN20-1a bearing 25 rDNA copies and rpa49Δ (25 rDNA copies [rpa49Δ]), and NOY1064 bearing 190 rDNA copies (190 rDNA copies [WT]). (right) Interpretive tracing of the genes is shown. Polymerases that appear on the gene are shown on the tracing by black dots. (B and C) Cumulative distribution function of the distances between adjacent polymerases in strains NOY1071 (red; 25 rDNA copies [WT]), BEN20-1a (blue; 25 rDNA copies [rpa49Δ]), and NOY1064 (green; 190 rDNA copies [WT]) is shown. Solid lines, empirical measurement; dashed lines, random (B) or simulated (C).

Figure 8.

Two successive elongating Pol I complexes allow a contact between Rpa49 and Rpa43 subunit of adjacent molecules. (A) Frequent association of adjacent Pol I depends on Rpa49. We generated diploid cells expressing two versions of Pol I containing an Rpa190 subunit fused to either the Myc or HA tag. These tagged proteins are efficiently detected in cell extract supernatants (SUP). After formaldehyde cross-linking and extensive DNase treatment, we immunoprecipitated Pol I using anti-HA antibodies (IP). We evaluated the amount of polymerase dimers/multimers by assessing the extent of Myc-tagged Rpa190 coprecipitation. We use a strain bearing only Myc-Tag Pol I as negative control (no HA tag; lanes 1 and 4). The presence of dimers/multimers is dependent on Rpa49, as shown by lack of Myc-tagged Rpa190 coimmunoprecipitation in rpa49 deletion background (rpa49Δ; compare lanes 5 and 6) (B) 3D model of Pol I as determined by cryo–electron microscopy. The positions of the Pol I–specific Rpa43/Rpa14 and Rpa49/Rpa34 heterodimers are shown in purple and in green, respectively. (C) Fitting of the atomic structure of the highly homologous Pol II elongation complex (Protein Data Bank accession no. 2YU9) into the cryo–electron microscopy structure of RNA Pol I. The transcribed DNA is shown in red, and the 10-subunit core Pol II is shown in blue. (D) Positioning of a second Pol I elongation complex (yellow) following a leading Pol I elongation complex (blue) by imposing a straight B-form linker DNA and a contact between the Pol I molecules. The black arrow represents the direction of RNA Pol I translation during elongation. (left) The model is rotated 180° around the vertical axis as compared with the right panel. (E) HA-tagged Rpa43 and Rpa34 bind to GST-tagged Rpa49 when expressed in E. coli. Expression and purification of recombinant GST-Rpa49 or GST alone in the presence of HA-tagged Rpa34 or Rpa43 is described in Materials and methods. The supernatants of E. coli cell lysates (SUP) were incubated with glutathione Sepharose 4B beads. After washing, GST-Rpa49 and GST were eluted with sample buffer (IP), and the fractions were analyzed by Western blotting to detect coprecipitation of HA-tagged Rpa34 (lanes 1–4) or Rpa43 (lanes 5–8). In the supernatant fractions, we detect full-length HA-Rpa34 and HA-Rpa43 and minor degradation products. Note that Rpa34 and Rpa43 appear specifically enriched in samples containing affinity-purified GST-Rpa49 compared with GST alone (compare lane 2 with lane 4 for Rpa34; compare lane 6 with lane 8 for Rpa43). White lines indicate that intervening lanes have been spliced out.