|Home | About | Journals | Submit | Contact Us | Français|
Upstream activating factor (UAF) is a multisubunit complex that functions in the activation of ribosomal DNA (rDNA) transcription by RNA polymerase I (Pol I). Cells lacking the Uaf30 subunit of UAF reduce the rRNA synthesis rate by ~70% compared to wild-type cells and produce rRNA using both Pol I and Pol II. Miller chromatin spreads demonstrated that even though there is an overall reduction in rRNA synthesis in uaf30 mutants, the active rDNA genes in such strains are overloaded with polymerases. This phenotype was specific to defects in Uaf30, as mutations in other UAF subunits resulted in a complete absence of rDNA genes with high or even modest Pol densities. The lack of Uaf30 prevented UAF from efficiently binding to the rDNA promoter in vivo, leading to an inability to activate a large number of rDNA genes. The relatively few genes that did become activated were highly transcribed, apparently to compensate for the reduced rRNA synthesis capacity. The results show that Uaf30p is a key targeting factor for the UAF complex that facilitates activation of a large proportion of rDNA genes in the tandem array.
A key early step of ribosome biogenesis is the transcription of ribosomal DNA (rDNA) by RNA polymerase I (Pol I). Transcription rates in the rDNA are high in exponentially growing cells and greatly reduced when nutrients are depleted (47). The rDNA in budding yeast (Saccharomyces cerevisiae) consists of a single tandem array on the right arm of chromosome XII with approximately 150 to 200 head-to-tail repeats that are each 9.1 kb long. Pol I-directed transcription of a repeat proceeds toward the centromere and produces the large 35S rRNA precursor (for a review, see reference 26). Each active yeast rDNA gene is transcribed by ~50 Pol I molecules, resulting in a Christmas tree-like structure observed in electron micrographs (EMs) of Miller chromatin spreads (12). Approximately 50% of the repeats in the array are used to achieve the high level of rRNA synthesis in log-phase cells (9). The mechanism by which cells initially open the rDNA chromatin structure for transcriptional activation, yet limit the number of active genes to approximately 50%, is unclear. When cells approach stationary phase, the histone deacetylase Rpd3p is required to return the rDNA genes to a closed chromatin structure (34).
Transcription of rDNA genes in yeast begins with the formation of a Pol I preinitiation complex at the promoter, and requires four major transcription factors: upstream activating factor (UAF), core factor (CF), TATA-binding protein (TBP), and Rrn3 (27, 28). UAF contains six subunits, i.e., Rrn5, Rrn9, Rrn10, Uaf30, and histones H3 and H4 (18, 19, 38), and directly associates with a region of rDNA located ~100 bp upstream of the transcription initiation site called the upstream element (UE). CF is a complex of three subunits, i.e., Rrn6, Rrn7, and Rrn11, that is analogous to mammalian SL1 (20-22) and centrally localizes to the core element of the promoter (19). TBP interacts with both UAF and CF, bridging the two factors (42). Rrn3 is the equivalent of mammalian TIF-IA (2) and forms an initiation-competent Rrn3-Pol I complex (48) that becomes depleted when cells are grown into stationary phase or have been treated with rapamycin, an inhibitor of TOR signaling that mimics nutrient deprivation (8, 25).
The UAF complex also affects the rDNA chromatin structure, preventing the rDNA genes from being transcribed by Pol II (46) and contributing to the SIR2-dependent silencing of Pol II-transcribed reporter genes in the rDNA (4, 7). The roles of individual UAF subunits in the initiation and silencing functions of this complex have remained largely unexplored. There are no structural data for any of the subunits except for histones H3 and H4, which, based on their function in the nucleosome, are likely to participate in DNA binding. Deleting the genes for the Rrn5p, Rrn9p, or Rrn10p subunit causes a loss of Pol I transcription but allows weak Pol II-mediated transcription of the rDNA genes (29). If the number of rDNA repeats then expands, these polymerase-switched (PSW) survivors can continue to grow using Pol II as the sole source of rRNA synthesis (29). Mutants lacking the Uaf30 subunit also allow the rDNA to be transcribed by Pol II, but most of the rRNA (~90%) continues to be produced by Pol I (38). The overall rRNA synthesis rate in uaf30Δ mutants is reduced by ~70%, but enough rRNA is made so that the cells survive.
In this study, we investigated the role of Uaf30 in activating Pol I transcription. In the absence of Uaf30, active rDNA genes were surprisingly overloaded with RNA Pol I molecules, a phenotype that is predicted to be associated with transcription elongation defects (35). In this case, however, the high Pol density appeared to be caused not by a problem with elongation but rather by an inability of the uaf30 mutants to activate normal numbers of rDNA genes in the tandem array. The few genes that were activated became heavily loaded with polymerases to compensate for the overall reduction in cellular rRNA production. Mutations in other UAF subunits, such as Rrn5 or Rrn9, which result in rDNA genes that are transcribed only by Pol II (46), produced “active” rDNA genes that were not overloaded with polymerases. Instead, the rDNA genes in the expanded array of these PSW strains were loaded with ~1 to 2 Pol II molecules each. Mechanistically, Uaf30 was shown by chromatin immunoprecipitation (ChIP) analysis to be required for efficient UAF association with the rDNA promoter in vivo. The UAF complex, therefore, has the potential to modulate the number of rDNA genes that are activated for Pol I transcription in response to nutrients.
Strains used in this study are listed in Table Table1.1. TAP-tagged strains were derived from BY4741 (Open Biosystems). Strains NOY886, NOY1051, and NOY1071, with defined rDNA copy numbers, were previously described (7, 12). All other strains were derived from the JB740 background previously used for rDNA silencing assays (4, 39). The uaf30::Tn3 insertion mutation was isolated from a genetic screen for Pol I transcription factors (R. Hontz and J. Smith, unpublished data). The Tn3 insertion consists of a promoterless lacZ cassette, LEU2, and the beta-lactamase ampicillin resistance gene (6). Yeast cells were grown in rich yeast extract-peptone-dextrose (YPD) or synthetic complete (SC) medium with components left out as indicated depending on the assay requirements (5). For 6-azauracil (6-AU) sensitivity tests, cells were patched onto SC medium without uracil (SC-ura) for plasmid selection and grown overnight. Cells were scraped from the overnight plate, resuspended in water, and adjusted to an optical density at 600 nm (OD600) of 1.0. Fivefold serial dilutions were spotted in 5-μl volumes onto SC-ura and SC-ura plus 6-AU.
Yeast were grown to mid-log phase in YPD medium supplemented with 1 M sorbitol. The YRH165 uaf30Δ mutant was sensitive to 1 M sorbitol, so 0.5 M was used for this strain. Miller spreads and EM with a JEOL 100 CX microscope were as previously described (12). The level of PSW transcription was determined by manually counting the number of polymerases and transcripts visualized along well-spread strands of the amplified rDNA array. To ensure that only rDNA repeats were measured, at least one of the transcripts on the strand needed to exhibit a terminal knob characteristic of the 35S genes (31). To determine the percentage of active genes within the PSW rDNA array, the distribution of polymerases was analyzed along chromatin strands that could be unambiguously traced for at least three repeat lengths (27.3 kb). This was to ensure that transcriptional activity was not overestimated by counting only active genes. To estimate the average polymerase density on active PSW rDNA genes, the minimum chromatin strand length was 9 kb (one repeat). The average number of polymerases/gene determined in this fashion is probably a slight underestimation due to the fact that a small percentage of the repeats in the tally are likely inactive.
Exponentially growing cells (40 ml in YPD) were pelleted, washed with cold water, and resuspended in 1.4 ml of cold Tris-EDTA. Half (0.7 ml) of the volume was transferred to a 24-well tissue culture plate, and 35 μl of a 200-μg/ml solution of 4,5′,8-trimethylpsoralen (Sigma) was added to each well and mixed. The cells were UV irradiated as previously described (34). Genomic DNA was isolated, digested with EcoRI, separated on a 1.3% agarose gel, and analyzed by Southern blotting using a 35S-specific rDNA probe labeled with [32P]dCTP by random priming. The hybridization signal was detected by autoradiography and quantitated using a Molecular Dynamics Storm phosphorimager and ImageQuant software.
A 1-ml aliquot of an overnight YPD yeast culture was pelleted and resuspended in 1 ml EDTA-Tris (0.05 M EDTA [pH 8], 0.01 M Tris-HCl [pH 7.5]). Cells were pelleted again, resuspended in 150 μl EDTA-Tris, and placed on ice. One at a time, the tubes were incubated for 2 min at 42°C with 1 μl Zymolyase solution (2 mg Zymolyase [20T] in 1 ml 10 mM sodium phosphate [pH 7.5]); 250 μl of 1.0% low-melting-point agarose was then added to each tube, gently resuspended, and cast into plugs that were allowed to solidify at 4°C for 1 h. The plugs were then incubated in 1 ml LET solution (0.5 M EDTA [pH 8], 0.01 M Tris-HCl [pH 7.5]) overnight at 37°C. Plugs were equilibrated at 4°C for 1 h. The LET solution was then replaced with 2 mg/ml proteinase K in NDS solution (0.5 M EDTA [pH 8], 0.01 M Tris-HCl [pH 7.5], 0.5 g Sarkosyl), and incubation was continued at 50°C for 2 days. The proteinase K-NDS solution was removed and replaced with 1 ml EDTA-Tris and incubated at 4°C for 1 h. Fresh EDTA-Tris was added and the plugs stored at 4°C. To digest chromosomes with BamHI, the plugs were preequilibrated in 500 μl 1× BamHI buffer (NEB) plus 0.1 mg/ml bovine serum albumin (BSA) on ice for 1 h. The buffer was replaced two times. Following the last incubation, 2 μl BamHI enzyme (NEB; 20,000 U/ml) was added, incubated on ice for 2 h, and then shifted to 37°C overnight.
Plugs were placed into the wells of a 1% agarose-0.5× Tris-borate-EDTA gel and run for 68 h on a Bio-Rad CHEF Mapper electrophoresis system. The running conditions were as follows: 120° angle, linear ramp (default), 3.0 V/cm, initial switch time of 300 s, and final switch time of 900 s; 0.5× Tris-borate-EDTA was circulated at 14°C as the running buffer. After completion, the gel was stained with 0.5 μg/ml ethidium bromide and destained, and a photograph was taken under UV light. The gel was enclosed in clear plastic wrap and exposed to UV light for 5 min in a UV-Stratalinker 2400 (Stratagene). The chromosomal fragments were then Southern transferred to a positively charged Immobilon nylon membrane (Millipore) in 10× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate). The rDNA array released by BamHI digestion was detected by hybridization with a 32P-labeled probe specific for the transcribed 35S region, followed by autoradiography.
ChIPs were performed as previously described (11), with some modifications. Yeast cells were grown in 10 ml YPD overnight and then reinoculated into 100 ml fresh YPD at an OD600 of 0.2 and shaken for 4 h at 30°C to an OD600 of ~1. These cultures were cross-linked with 1.35 ml formaldehyde (37%, wt/wt) for 20 min. Harvested cells were washed with 10 ml cold TBS (300 mM NaCl, 40 mM Tris [pH 7.5]) and flash frozen. Thawed cells were resuspended in 0.6 ml FA-lysis 140 buffer plus protease inhibitors (50 mM HEPES, 140 mM NaCl, 1.0% Triton X-100, 1.0 mM EDTA, 0.1% sodium deoxycholate, 250 μl 10× protease inhibitor cocktail [Sigma], 24 μl 100 mM benzamidine, 6 μl fresh 500 mM phenylmethylsulfonyl fluoride). Cells were disrupted in a FastPrep FP120 machine (Bio101/ThermoSavant) at 4°C. Cell extracts were sonicated with eight 10-s pulses (30% output, 90% duty cycle) on ice and then centrifuged at 14,000 rpm in an Eppendorf microcentrifuge for 5 min at 4°C. The protein content in the supernatants was quantified using a Bradford assay (Bio-Rad). For immunoprecipitations, 1.25 mg of protein for each extract was incubated in a 0.6-ml microcentrifuge tube with FA-lysis 140 solution in a total volume of 400 μl. A 1:50,000 dilution of anti-protein A antibody (P-3775; Sigma) was added and rotated overnight at 4°C. One-tenth of the chromatin extract volume used for the immunoprecipitation was used as the input control extract. The tubes were centrifuged at 14,000 rpm for 2 min at 4°C, and the supernatant was added to a new 0.6-ml microcentrifuge tube containing 60 μl of protein A-Sepharose beads (50% slurry), and rotated at 4°C for 4 h. Beads were spun at 5,000 rpm for 30 s at room temperature and then washed four times with 0.5 ml FA-lysis 140 buffer, four times with 0.5 ml FA-lysis 500 buffer (50 mM HEPES, 500 mM NaCl, 1.0% Triton X-100, 1.0 mM EDTA, 0.1% sodium deoxycholate), and four times with 0.5 ml LiCl detergent wash buffer (10 mM Tris-HCl [pH 8], 250 mM LiCl, 0.5% NP-40, 0.5% sodium deoxycholate, 1.0 mM EDTA). DNA and protein were eluted two times with 75 μl of 5× Tris-EDTA plus 1% sodium dodecyl sulfate (SDS). The 150-μl total volume was incubated overnight at 65°C to reverse cross-links. DNA was purified using a Qiagen PCR purification kit. PCR conditions were as follows: 92°C for 3 min (1 cycle); 92°C for 30 s, 52°C for 30 s, and 72°C 45 s (22 cycles for rDNA or 29 cycles for ACT1); and 72°C for 5 min (1 cycle). Reaction products were separated on a 1.5% agarose gel and stained with ethidium bromide. Band intensities were quantitated using ImageQuant software (Molecular Dynamics). Oligonucleotides used for PCR are listed in Table Table22.
A linear fragment from plasmid pNOY3238 (44), which contains the rRNA gene promoter (position −210 to +124 with respect to the transcription start site), was used as a probe for footprinting experiments. The top strand was footprinted by linearizing the plasmid with BstBI (position +24) and radioactively labeling by incubation with Exo− Klenow (New England Biolabs) and 100 μCi [α-32P]dCTP (NEN-Perkin-Elmer; 6000 Ci/mmol) for 55 min at 37°C, followed by the addition of deoxynucleoside triphosphate mix and incubation for 15 min. The DNA was recovered by ethanol precipitation. The linearized, labeled plasmid was digested with HindIII (position −234), and the 258-bp fragment containing the promoter was isolated on a 5% native polyacrylamide gel and purified using Elutip D (Schleicher and Schuell) as described by the manufacturer. The DNA was then precipitated with ethanol and resuspended in water.
Purification of UAF from wild-type (WT) and uaf30Δ strains has been described previously (38). UAF was diluted in UAF dilution buffer (20 mM Tris-acetate [pH 7.9], 1 mM dithiothreitol, 0.2 mg/ml acetylated BSA, and 20% glycerol) and incubated with the labeled DNA fragment in DNA binding buffer (40 mM Tris-HCl [pH 7.9], 100 mM KCl, 10 mM MgCl2, and 100 μg/ml acetylated BSA) in a 50-μl reaction mixture. After 20 min at room temperature, 1 U of RQ1 DNase I (Promega) diluted fivefold in UAF dilution buffer was added. The reaction was stopped after 2 to 3 min by the addition of 100 μl of DNase stop mix (0.45 M sodium acetate [pH 5.2] and 15 mM EDTA). An equal volume of phenol-chloroform-isoamyl alcohol was then added to the reaction mixture. The DNA in the aqueous layer was precipitated with ethanol in the presence of 10 mg glycogen. The precipitated DNA was washed once with cold 70% ethanol, dried, and resuspended in loading dye (95% formamide, 20 mM EDTA, and bromophenol blue and xylene cyanol blue dyes). The reaction products were analyzed on a 7% acrylamide-7 M urea denaturing gel. Sequencing reaction mixtures with plasmid pNOY3238 using dideoxynucleotides and a labeled primer that corresponds to the BstBI site (5′-CGAACTTGTCTTCAACTGCTT) were also loaded on the same gel. The gel was dried and analyzed by autoradiography.
Log-phase yeast cultures (5 ml) were pelleted and then extracted in 0.5 ml 20% trichloroacetic acid by vortexing with glass beads. Lysates were transferred to a new microcentrifuge tube. The glass beads were washed with 0.5 ml 5% trichloroacetic acid, which was then combined with the first lysate. The tubes were spun at 3,000 rpm in an Eppendorf 5810R microcentrifuge for 10 min at 4°C and the supernatant discarded. Pellets were resuspended in 200 μl sample buffer (50 mM Tris-HCl [pH 6.8], 2% SDS, 10% glycerol, 0.1% bromophenol blue), 10 μl β-mercaptoethanol, and 50 μl 2 M Tris base, and 10-μl aliquots of the samples were separated on 10% SDS-polyacrylamide gels and transferred to an Immobilon-P membrane (Millipore). The membrane was blocked with 5% milk in phosphate-buffered saline-0.1% Tween 20 and then probed with a 1:50,000 dilution of anti-protein A primary antibody (P-3775; Sigma) or a 1:5,000 dilution of antitubulin monoclonal antibody (B-5-1-2; Sigma) in 25 ml phosphate-buffered saline-0.1% Tween for 1 h and then with a 1:5,000 dilution of goat anti-rabbit or anti-mouse secondary antibody conjugated to horseradish peroxidase. Proteins were detected by chemiluminescence using ECL (GE Healthcare). ImageQuant software was used to quantify bands from digitally scanned films.
Strains with UAF30 deleted have an ~70% reduction in the rRNA synthesis rate and permit transcription of rDNA genes by both Pol I and Pol II, with Pol II being responsible for ~10% of the rRNAs synthesized (38). The specific role that Uaf30 plays in the UAF complex to facilitate transcription of rDNA genes has remained largely unexplored. To gain insight into its function, we visualized the effects of two different uaf30 mutants on rDNA gene transcription at the single-gene level using EM of Miller chromatin spreads. One mutation was a transposon insertion (uaf30::Tn3), and the other was a deletion of the entire open reading frame (uaf30Δ). As shown in Fig. Fig.1A,1A, active rDNA genes in the uaf30 mutants were greatly overloaded with polymerases (>100/gene) compared to those in the WT parent strain (~49/gene). In contrast, treatment of WT cells with rapamycin for 1 h, which reduces Pol I initiation (8, 45), caused a decrease in the number of polymerases loaded. Since UAF had previously been implicated in stimulating transcription initiation (1, 19) and a uaf30Δ mutation reduces the transcription level by 70% (38), the high polymerase density for the uaf30 mutants was very unexpected. The high density was more similar to the phenotype of a top1Δ mutant, which also produced active genes overloaded with polymerases (Fig. (Fig.1A).1A). Cells deficient in topoisomerase I were previously shown to be defective in Pol I elongation (37), and a transcription defect in which the elongation rate is reduced relative to the initiation rate should lead to more polymerases on the gene, as we observed.
Another possible explanation for the high polymerase density of rDNA genes in the uaf30 mutants is that the size of the tandem array was reduced. The transcription level of individual rDNA genes can be inversely correlated with the repeat copy number of the tandem array, such that rDNA genes within a small tandem array (42 copies) are more heavily loaded with polymerases than genes in a larger array (143 copies) (12). The array sizes of the uaf30 mutants and various control strains were, therefore, determined by pulsed-field gel analysis (Fig. (Fig.1B).1B). The array size of the WT parental strain, YRH4, was significantly greater than that of the 143-copy control strain and was estimated as ~175 repeats. The uaf30Δ mutant array was approximately equal in size to that of YRH4, while the uaf30::Tn3 mutant array was larger (~200 repeats), similar to previous results for a different uaf30 deletion strain (38). To demonstrate that the rDNA array in the YRH4 background was capable of being reduced in size, we deleted the SPT4 gene, which was earlier shown to shorten the array and induce polymerase-dense rDNA genes in a different strain background (35). Spt4 functions as an elongation factor for both Pol I and Pol II transcription (14, 35). As shown in Fig. Fig.1C,1C, the spt4Δ mutant in the YRH4 background (YRH245) did have a smaller array. These results indicated that the high density of polymerases in the uaf30 mutants was not caused by a simple decrease in the number of rDNA genes.
Since the rDNA array size of the uaf30 mutants was not less than normal, it was difficult to reconcile the high polymerase densities with their overall transcription defect without invoking a possible elongation defect. However, when analyzing Miller spreads of the uaf30 mutants, it was hard to find an active repeat with its immediate neighboring genes also active (~5%, compared to ~40% for a 143-copy strain), suggesting that there might be a reduction in the number of active genes in these mutants. Indeed, examination of the chromatin spreads at lower magnifications indicated that there were relatively few active genes in the spread nucleoli of uaf30Δ mutants compared to the WT strain (Fig. (Fig.2A).2A). This reduction was confirmed by gauging the proportion of genes within the tandem array having an open chromatin structure by using a psoralen photo-cross-linking method, which differentiates and separates the active and inactive rDNA genes on an agarose gel (9). As shown in Fig. 2B and C, the proportion of open rDNA genes in both uaf30 mutants was reduced in comparison to that in the WT YRH4 parental strain. The reason for the generally increased gel mobility for the rDNA bands in the uaf30Δ mutant is unknown, but it could be either related to an alteration in chromatin structure that affects cross-linking efficiency or a sample preparation artifact, as a similar shift was observed with a PSW strain described below (see Fig. Fig.4B4B).
The fact that so few rDNA genes were open/active for Pol I transcription in the uaf30 mutants raised the question of whether the overall cellular amount of Pol I had been altered. Steady-state Western blot analysis of TAP-tagged Rpa135, the second-largest subunit of Pol I, indicated there was no consistent difference in the amount of this Pol I subunit in various uaf30Δ isolates compared to the WT parent (Fig. (Fig.3A).3A). One WT isolate had an approximately twofold-higher Rpa135 protein content, most likely due to a gene duplication, and was not used for ChIP experiments. Despite normal protein levels, overall Pol I association with the transcribed region of the rDNA gene was greatly reduced in the uaf30Δ mutant compared to the WT as measured by ChIP for the TAP-tagged Rpa135 (Fig. (Fig.3B).3B). This result was fully consistent with the Miller spread and psoralen cross-linking data showing few rDNA genes being activated for Pol I transcription. Additionally, the data suggested that Pol I was not significantly loaded onto rDNA genes other than those displaying the high Pol density. Taken together, these results strongly suggest that the high Pol I density in uaf30 mutants is caused by a reduction in the number of genes that are being transcribed by Pol I. We hypothesize that the cell attempts to compensate for the very low number of active genes by increasing the number of polymerases transcribing those genes but that it is unable to restore the overall rRNA synthesis level. In contrast, mutations in topoisomerase I (TOP1) result in genes densely packed with polymerases due to a transcription elongation defect that is not related to a reduction in the number of active genes. Although genes were overloaded with polymerases (Fig. (Fig.1A),1A), the nucleoli of a top1Δ mutant still had an apparently typical number of active genes as determined by chromatin spreading (see Fig. S1 in the supplemental material).
Based on the EM data, we estimated the number of active rDNA genes in the uaf30Δ mutant using the following calculation. If the elongation rate in two different strains is the same, then the difference in the rate of steady-state rRNA synthesis should be reflected by the difference in the total number of Pol I molecules engaged in transcription per cell between the two strains. Thus, since the rate of rRNA synthesis in the uaf30Δ mutant is ~30% of that in control WT cells (38), the total number of Pol I molecules engaged in transcription in the mutant cells, i.e., the Pol I density per active gene (~100 [Fig. [Fig.1A])1A]) multiplied by x (the number of active genes per cell), should equal 0.3 multiplied by the product of the Pol I density per active WT gene (~50 [Fig. [Fig.1A])1A]) and the number of active genes per WT cell (~85 for YRH4; 50% of the estimated 175 repeats). Based on this calculation [100(x) = 0.3(50)(85)], the number of active genes (x) in the uaf30Δ mutant is ~13. This estimate correlates well with the low-magnification EM image of the representative uaf30Δ nucleolar region shown in Fig. Fig.2A2A and the 12 to 16 active genes that were observed in three other well-spread nucleoli of YRH165 (data not shown).
Approximately 10% of the rRNA synthesis in a uaf30Δ mutant is derived from Pol II and 90% from Pol I (38). Strains lacking the Rrn5 or Rrn9 subunit of UAF are unable to utilize Pol I but do allow the rDNA to be transcribed at a reduced level by Pol II. Viable rrn5Δ or rrn9Δ variants can be obtained that have undergone spontaneous expansion of the tandem array size through amplification of the rDNA gene copy number (up to ~400 copies), allowing sufficient rRNA synthesis for growth (29). These types of variants are called PSW strains. It has remained unclear whether all the rDNA genes in these expanded tandem arrays are weakly transcribed by Pol II or whether only a limited number of the genes become significantly transcribed (as in uaf30 mutants with Pol I), leaving most other copies completely inactive. To distinguish between these possibilities and for comparison with uaf30 mutants, we performed Miller spread analysis on an rrn9Δ rpa135Δ PSW strain (NOY897) that can utilize only Pol II for rRNA synthesis (Fig. (Fig.4).4). Analysis of dispersed chromatin from >400 cells revealed an absence of normal nucleolar regions. Instead, large expanses of chromatin displaying a very low density of rRNA transcripts were observed (Fig. (Fig.4A,4A, top); for comparison, see nucleoli from the WT strain (Fig. (Fig.2A)2A) and the top1Δ mutant (see Fig. S1 in the supplemental material). rRNA transcripts are easily distinguished from mRNA transcripts by their greater length and by their characteristic terminal knobs, which correspond to SSU processomes (31). Examples of such transcripts are shown at higher magnification in Fig. Fig.4A4A (bottom) and in Fig. Fig.5A.5A. No typical rDNA genes exhibiting the distinct “Christmas tree” morphology of highly active genes were observed. The size of the rDNA array in the PSW strain was previously determined to be ~400 repeats (29), which nearly doubles the size of chromosome XII. Consistent with the expanded array size, chromatin areas that displayed sparse rRNA transcripts in the PSW strain were very large (Fig. (Fig.4A).4A). In WT cells, however, comparatively few nascent transcripts are seen outside the nucleolar region.
In the PSW strain, the psoralen cross-linking assay failed to detect any discernible bands in the Southern blot that corresponded to the gel region where rDNA genes with an open, psoralen-accessible structure migrate (Fig. (Fig.4B).4B). Higher bands in the gel for both the WT and PSW strains are caused by incomplete EcoRI digestion and do not correspond to the open genes. This result is consistent with previous findings that Pol II transcription of rDNA on a plasmid also did not cause increased accessibility to psoralen (10). Therefore, we used EM to determine the percentage of active genes within the rDNA array of the PSW strain. In total, 1,378 kb along 36 different chromatin strands were analyzed, representing 151 repeats, of which 127 (84%) were active. The average polymerase density on the “active” Pol II-transcribed rDNA genes in the PSW strain was determined by counting polymerases along 3,403 kb of chromatin (374 repeats). The total polymerase count was 590, yielding an average of 1.6 polymerases/gene. Therefore, each PSW cell has ~538 Pol II molecules engaged in making rRNA (~336 active genes with 1.6 polymerases/gene), while normal yeast cells have ~3,750 Pol I molecules engaged in making rRNA (~75 active genes with ~50 polymerases/gene (12). Thus, EM analysis indicates that PSW cells have ~14% as many polymerases transcribing rRNA as normal cells, in agreement with RNA labeling experiments showing that the rRNA accumulation rate in PSW strains is about 10 to 20% of that of control strains (data not shown). These results are consistent with a model for PSW formation in which a sufficiently large array can compensate for a very low level of rRNA production from each individual repeat. In contrast, specific loss of the Uaf30 subunit of UAF results in high Pol I density on a small number of genes.
Since uaf30 mutants produce rRNA using both Pol I (90%) and Pol II (10%), we addressed the possibility that Pol II transcription of the rDNA genes was also contributing to the high polymerase density. If Pol II molecules do not elongate on the rDNA as efficiently as Pol I, then the intrusion of Pol II could hinder Pol I elongation, resulting in a polymerase-dense phenotype and reduced levels of transcription. To test this possibility, we utilized a deletion of the RPD3 gene, which encodes a class I histone deacetylase. Deletion of RPD3 strengthens the silencing of Pol II-transcribed reporter genes in the rDNA (40, 43) and inhibits the Pol II-mediated synthesis of rRNA that occurs in uaf30Δ or rrn9Δ mutants (30). Miller spread analysis of active rDNA genes from a uaf30::Tn3 rpd3Δ double mutant clearly demonstrated that there was no significant difference in polymerase densities between the uaf30 mutants and the double mutant (Fig. (Fig.5A5A shows results for representative genes). This result suggests that high polymerase densities were not caused by Pol II clogging transcription of the rDNA. Furthermore, it suggests that the Pol II-produced rRNA in uaf30 mutants likely comes from repeats that are not transcribed by Pol I but are weakly transcribed by Pol II, similar to the rDNA genes in PSW strains. Such genes are occasionally seen in uaf30 chromatin spreads (data not shown).
As mentioned above, a defect in Pol I elongation could potentially contribute to uaf30 mutants producing rDNA genes with high polymerase densities. One of the common means of testing whether a yeast mutant has a transcription elongation defect is growing cells on media containing 6-AU, which slows the growth of yeast mutants defective in Pol I or Pol II elongation (36). As shown in Fig. Fig.5B,5B, ,6-AU6-AU did slow the growth of a uaf30::Tn3 mutant and the top1::Tn3 positive control. However, we hypothesized that rather than this being a consequence of an elongation defect, 6-AU might be detrimental to uaf30 mutants and other cells with a very limited number of active genes. Any slowing of rRNA production caused by 6-AU in such cases could tip the balance between normal and poor growth. To test this idea, we utilized a set of strains with stable array sizes of 143, 42, or 25 rDNA gene copies (7, 12) but with intact UAF30 genes. The 25-copy strain (25c) was sensitive to 6-AU, while the 42- and 143-copy strains were resistant (Fig. (Fig.5B,5B, bottom). Even though rDNA genes are dense in both the 25-copy and 42-copy strains, the 42-copy strain produces enough rRNA to withstand the reduction in elongation caused by the 6-AU treatment. Since the 25-copy strain is sensitive to 6-AU, it is not surprising that the uaf30Δ strain with fewer than 25 active rDNA genes per cell is also sensitive. Consistent with this hypothesis, crossing a uaf30Δ mutant with a top1Δ mutant (having an elongation defect) resulted in a severe synthetic slow-growth phenotype in double mutant haploid spores that were produced from tetrad dissections of the a heterozygous diploid (Fig. (Fig.5C).5C). Cumulatively, the results described so far strongly suggest that the transcription defect of a uaf30 mutant is due to an inability to activate a large number of rDNA genes for Pol I transcription and that an elongation effect, if any, is minimal.
We next tested whether a lack of Uaf30 altered the association of UAF with the rDNA promoter in vitro by using a DNase I footprinting assay. UAF complex purified from a uaf30Δ strain retained approximately 50% of its in vitro Pol I transcription stimulatory activity on naked DNA templates, but a direct effect on binding had not been tested (38). Importantly, all other UAF components were present in this mutant preparation (38). For the WT UAF complex, the major region of DNase I footprinting protection on the noncoding strand occurred between nucleotides −48 and −107 of the promoter UE, with strong hypersensitive sites at nucleotides −113/−110 and −77/−76 (Fig. (Fig.6A,6A, lanes 2 to 4). For the mutant complex, the most obvious difference was loss of the −113/−110 hypersensitive site and a loss of the protection from −100/−97 to −107 (Fig. (Fig.6A,6A, lanes 5 and 6). A similar loss of protection was observed with the coding strand (data not shown), indicating that the presence of Uaf30 in the UAF complex extends contacts with the DNA template in the upstream direction by ~10 to 15 bp on both strands. Other regions of protection were affected less than twofold, meaning that the mutant complex still efficiently binds the promoter in the context of naked DNA, but its footprint is altered.
Since uaf30 mutants were unable to activate normal numbers of rDNA genes in the rDNA array, we hypothesized that the Uaf30 subunit of UAF played a key role in promoter binding in vivo. To test this possibility, we measured the binding of a TAP-tagged version of the Rrn5 UAF subunit to the rDNA promoter in NTS2 using ChIP. As expected, in a WT strain background Rrn5-TAP was efficiently bound to NTS2 but not to the transcribed 25S region (Fig. (Fig.6B).6B). Due to random chromatin shearing, a small amount of binding was detected at the 5′-ETS region, which is immediately adjacent to the promoter. In contrast, binding of Rrn5-TAP to NTS2 or the adjoining 5′-ETS region was surprisingly undetectable in the uaf30Δ mutant (Fig. (Fig.6B).6B). As a control, we tested the binding of a TAP-tagged version of the Rrn7 CF subunit. Rrn7-TAP binding to NTS2 was also decreased in the uaf30Δ mutant (Fig. (Fig.6B),6B), but only by a level similar to the reduction of Rpa135-TAP binding observed in the uaf30Δ mutant (Fig. (Fig.3B),3B), which was most likely a consequence of the low number of active rDNA genes. Importantly, TAP-tagged Rrn5 and Rrn7 were efficiently expressed in the uaf30Δ mutant (Fig. (Fig.6C).6C). Rrn5-TAP was slightly decreased in the mutant but not enough to explain the dramatic loss of binding to the promoter. These results strongly suggest that Uaf30 is important for proper and stable UAF association with the promoter. Coinciding with the decreased number of rDNA genes that are bound and activated by the mutant UAF complex, recruitment of CF (Rrn7) and Pol I (Rpa135) is restricted to the genes that remain in the active chromatin state.
The goal of this study was to provide new insights into the mechanism by which the Uaf30 subunit of the Pol I transcription factor UAF contributes to the activation of rDNA transcription. Our initial finding that the active rDNA genes in uaf30 mutants were overloaded with RNA polymerases was unexpected, since all previous data had pointed to the UAF complex being involved in the transcription initiation step (1, 19). Our ChIP experiments showed that in the absence of Uaf30, the UAF complex does not efficiently associate with the rDNA promoter in vivo, and as a result, very few genes become activated. The relatively few rDNA genes that the cell does activate become heavily loaded with polymerases, apparently in an attempt to compensate for the overall reduction in rRNA synthesis capacity. Without the ability to efficiently activate genes, the cell is able to synthesize rRNA at only ~30% of its normal rate (38).
Although we cannot conclusively rule out elongation defects as a contributing factor to the high polymerase density of active rDNA genes in the uaf30 mutants, if such a defect exists, several lines of evidence strongly suggest that elongation does not have a major impact. First, the 6-AU sensitivity of the uaf30::Tn3 mutant can be accounted for by a low number of active genes, which makes low-copy-number UAF30+ strains sensitive to 6-AU. Second, the UAF complex does not significantly associate with the transcribed region of the rDNA according to the ChIP data. Some elongation factors, such as Spt4/Spt5, do associate with the transcribed region (35). Third, only about 13 rDNA genes are active in a typical uaf30Δ mutant. If the elongation rate in the mutant was significantly less than that in the WT, the number of active genes in the mutant would have to be much higher than this value, which our observations do not support. The decrease in the number of active genes appears to be the major cause for the decrease in overall rRNA synthesis in the uaf30 mutants.
One of the surprises from this study was the lack of detectable in vivo UAF binding to the rDNA promoter in the uaf30 mutants, despite the ability of the purified mutant complex to associate with the promoter DNA in vitro. Even though we were unable to detect the TAP-tagged Rrn5 subunit at the promoter by ChIP in the uaf30Δ mutant, we hypothesize that the mutant complex likely does contact the promoter of the small number of active genes at some point. The mutant complex could take on an altered structure that affects its residence time on the promoter in the context of chromatin. This would reduce the probability of forming the initial preinitiation complex. In contrast, when the Rrn5 or Rrn9 subunit of UAF is missing, we do not detect any rDNA genes with active Pol I transcription or even dense Pol II transcription. This is also consistent with the less defective uaf30 mutant complex retaining some capacity for activation in vivo. Once the chromatin has been opened by the mutant UAF complex and transcription initiated at one of the rare active genes, the excess of other transcription components such as CF, TBP, Rrn3, and Pol I could then allow the cell to perform multiple rounds of transcription despite the poor association of mutant UAF. We also cannot currently rule out the possibility that an altered UAF structure reduces the ability to detect the tagged Rrn5 subunit by ChIP, which, coupled with its presence on only a few genes, would reduce its ChIP signal even further.
Based on earlier studies with rrn9Δ PSW and uaf30Δ strains, a model was proposed in which Pol I and Pol II transcription of the rDNA genes were reciprocal and mutually exclusive (7, 30). Consistent with that model, deleting the RPD3 gene, which prevents Pol II-mediated transcription of the rDNA genes (30), had no effect on either normal polymerase densities in a UAF30 background (34) or the high polymerase densities of a uaf30 mutant (Fig. (Fig.5A).5A). Furthermore, Miller spreads of the rrn9Δ PSW strain demonstrated that Pol II is not sufficient to generate normal polymerase densities, producing an average of only one to two polymerases on each rDNA gene in such mutants (Fig. (Fig.4).4). Therefore, we propose that the “open” genes observed in psoralen cross-linking assays and Miller spreads in uaf30 mutants are transcribed exclusively by Pol I, while the Pol II-derived rRNAs are produced from the “closed” genes in both the uaf30 and PSW mutants. Given the reciprocal nature of the two polymerases on the rDNA promoter, it remains possible that the loss of silencing and intrusion of Pol II in the uaf30Δ strains could contribute to the dramatic reduction in the number of open, Pol I-transcribed rDNA genes by inhibiting formation of the Pol I preinitiation complex.
When the UAF complex lacks Uaf30, its DNase I footprint on the promoter UE is smaller by ~10 to 15 bp, but overall binding to template DNA in vitro is affected less than twofold (Fig. (Fig.6A).6A). We do not yet know whether Uaf30 makes direct contacts with the DNA template or whether the lack of Uaf30 simply alters the local structure of the DNA binding region of the complex such that the footprint of another subunit is altered. In the context of a chromatin template in vivo, Uaf30 is critical for the stable association of UAF with the rDNA. It is possible that the change in footprinting observed in vitro for the mutant complex has a significant effect on binding stability in vivo. Uaf30 could also function in remodeling the in vivo rDNA chromatin in a way that facilitates stable binding of UAF to the promoter. Uaf30 has a conserved SWIB domain that is shared with the BAF60a subunit of the human SWI/SNF chromatin-remodeling complex (38). While the exact function of the BAF60a subunit has not been determined, it is known to interact with multiple transcriptional regulatory proteins, including steroid hormone receptors and Fos/Jun (15, 16). Uaf30 could potentially provide a binding surface that recruits to the rDNA promoter unidentified transcriptional coactivators and/or chromatin remodelers that enhance stable complex formation. Alternatively, UAF could have its own intrinsic chromatin-remodeling activity that is specific for the rDNA.
By targeting UAF to the promoter, Uaf30 facilitates activation of the large proportion (~50%) of rDNA genes that are normally turned on during rapid cellular growth. This result makes UAF a potential regulatory point for the cell to control the number of active genes in response to environmental changes. Outside of histones H3 and H4, posttranslational modifications of UAF subunits in yeast have not been reported. H3 and H4 are components of UAF (18) and are well known to be the targets of multiple posttranslational modifications in the context of nucleosomes. However, it is not known if they are modified in the context of UAF. Since UAF is bound to the rDNA promoter of both active and inactive repeats in WT cells (24) and even remains associated with the rDNA promoter during stationary phase (8), posttranslational modification of H3, H4, or one of the other UAF subunits could be a potential mechanism for controlling the activation of genes in response to altered nutrient levels through modulating UAF function (18, 34). Analogously, the negative impact that deleting UAF30 has on UAF function results in a decrease in the number of active rDNA genes.
Homologs of Uaf30 and the other UAF subunits have not been identified in vertebrates, which instead have an HMG box-containing protein called upstream binding factor (UBF) that binds to the rDNA promoter and stimulates rRNA synthesis (17). In this sense, UBF has some functional similarity to UAF, although UBF functions as a dimer (23) and not in a multisubunit complex. Another key difference is that in addition to binding the promoter, UBF also binds within the rDNA genes (32) and has been reported to function in Pol I elongation (41). The activity of UBF is also modulated by phosphorylation and acetylation (33). Hmo1 is an HMG box protein found in S. cerevisiae that promotes Pol I transcription and has been proposed to have some functional similarity to UBF (13). The factors that function analogously to UAF in vertebrates to ensure activation of a large proportion of rDNA genes remain unknown. UBF is formally a candidate, but it is also possible that a yet-unidentified factor analogous to UAF exists in vertebrates (46).
We thank the labs of D. T. Auble and M. W. Mayo for access to their PCR machines and antibodies, Mary Bryk for technical advice, Cathleen Josaitis for participation in earlier UAF footprinting with wild-type yeast, and Dan Burke for providing yeast strains. We also thank Yvonne Osheim for critical reading of the manuscript and helpful discussions. Technical assistance with the EM work was provided by Martha Sikes.
This work was supported by NIH grant GM61692 and American Heart Association grants 055490U and 0755633U to J.S.S., NIH grant GM63952 to A.L.B., and NIH grant GM35949 to M.N.
Published ahead of print on 2 September 2008.
†Supplemental material for this article may be found at http://mcb.asm.org/.