The C-terminal region of Net1 is an activator of RNA polymerase I transcription with conserved features from yeast to human

RNA polymerase I (Pol I) synthesizes ribosomal RNA (rRNA) in all eukaryotes, accounting for the major part of transcriptional activity in proliferating cells. Although basal Pol I transcription factors have been characterized in diverse organisms, the molecular basis of the robust rRNA production in vivo remains largely unknown. In S. cerevisiae, the multifunctional Net1 protein was reported to stimulate Pol I transcription. We found that the Pol I-stimulating function can be attributed to the very C-terminal region (CTR) of Net1. The CTR was required for normal cell growth and Pol I recruitment to rRNA genes in vivo and sufficient to promote Pol I transcription in vitro. Similarity with the acidic tail region of mammalian Pol I transcription factor UBF, which could partly functionally substitute for the CTR, suggests conserved roles for CTR-like domains in Pol I transcription from yeast to human.


Author summary
The production of ribosomes, cellular factories of protein synthesis, is an essential process driving proliferation and cell growth. Ribosome biogenesis is controlled at the level of synthesis of its components, ribosomal proteins and ribosomal RNA. In eukaryotes, RNA polymerase I is dedicated to transcribe the ribosomal RNA. RNA polymerase I has been identified as a potential target for cell proliferation inhibition. Here we describe the C-terminal region of Net1 as an activator of RNA polymerase I transcription in baker's yeast. In the absence of this activator RNA polymerase I transcription is downregulated and cell proliferation is strongly impaired. Strikingly, this activator might be conserved in human cells, which points to a general mechanism. Our discovery will help to gain a better understanding of the molecular basis of ribosomal RNA synthesis and may have implications in developing strategies to control cellular growth. Introduction RNA polymerase I (Pol I) transcribes the precursor for three out of the four ribosomal RNAs (rRNAs), which are essential components of ribosomes, required for cell growth and proliferation. In proliferating cells of S. cerevisiae (hereafter called yeast), Pol I activity accounts for over 50% of the cellular transcriptional activity [1]. Robust rRNA production is supported by repetitive gene arrays at one or multiple chromosomal locations, collectively called the ribosomal DNA (rDNA) loci (see Fig 1 upper panel, yeast rDNA). Furthermore, the specialized Pol I transcription machinery, including dedicated transcription factors, promotes efficient rRNA synthesis (reviewed in [2,3]). In the past, many factors supporting Pol I transcription in vivo and in vitro have been identified in various organisms. However, it is still an open question how the observed high transcriptional output is mechanistically achieved. Basal mechanisms of Pol I transcription might be conserved since there appears to be a significant degree of similarity even between distantly related species. Both, mammalian and yeast promoters are arranged in two main domains including a core element (CE, Fig 1A), required for transcription initiation in vitro, and an upstream (control) element (UE in yeast (Fig 1A), UCE in mammals), supporting activated transcription under certain conditions (reviewed in [3]). Whereas promoter DNA sequences are unrelated, components of CE-binding complexes, the yeast core factor (CF) and the human Selectivity Factor 1 (SL1) share functional and structural homology [4][5][6][7]. CF and SL1 are required to recruit the conserved initiation competent Rrn3-Pol I complex to the rDNA promoter [8,9]. Stable recruitment of CF to the rDNA promoter in yeast depends on the 6-subunit upstream activation factor (UAF) binding to the UE [10]. In analogy, the high mobility group (HMG)-box protein upstream binding factor (UBF) interacts with the UCE and stabilizes the SL1 complex at mammalian rDNA promoters [11,12]. UBF has, however, no reported homology with UAF components. Apart its function in Pol I pre-initiation complex (PIC) formation, UBF can bind the entire rDNA region transcribed by Pol I presumably to maintain an open chromatin structure [13,14]. In yeast, the HMGbox protein Hmo1 is a component of open rDNA chromatin [15,16], and genetic data suggests functional conservation between distinct HMG-boxes in human UBF and Hmo1 [17].
Regulation of Pol I transcription by post-translational covalent modifications of proteins has been extensively described and characterized in mammals (reviewed in [18,19]). Thus, the polymerase, Rrn3, SL1 and UBF are targets of phosphorylation or acetylation, and many factors either attaching or removing specific modifications marks have been reported. Post-translational modifications occur in response to intra-or extracellular signals mainly implicated in the regulation of cell growth and proliferation. Like in higher eukaryotes, yeast ribosome biogenesis responds to different growth conditions (reviewed in [20]). However, for this organism detailed knowledge about post-translational modifications regulating the activity of the basal Pol I transcription machinery is rather limited. Recently, yeast phosphatase Cdc14, the orthologue of mammalian Cdc14B, has been proposed to downregulate Pol I transcription by dephosphorylation of a polymerase subunit in late anaphase of the cell cycle [21]. This differs from observations in higher eukaryotes, where Cdc14B phosphatase activity is required for reactivation of Pol I transcription at the end of mitosis [22]. Likewise, both, the yeast NADdependent lysine deacetylase Sir2, and its mammalian homologue SIRT1 have been differentially implicated in the regulation of rDNA transcription. While the mammalian SIRT1 protein deacetylates components of the Pol I transcription machinery [23], yeast Sir2 rather targets RNA polymerase II (Pol II) transcription in rDNA repeats likely by acting on acetylated histones [24,25]. Therefore, although post-translational covalent modifications by homologous factors might influence rDNA transcription in yeast and mammals, it is unclear if the regulation by the respective modification marks is conserved. In the middle: enlargement of one rDNA repeat depicting 35S rDNA transcription units, intergenic sequences 1 and 2 (IGS1,2), the 5S rDNA, an autonomous replication sequence (ARS), and a bi-directional Pol II-dependent promoter (E-pro). Positions of restriction sites (E, EcoRI; X, XcmI) used in ChEC and psoralen crosslinking experiments are depicted. At the bottom: cis elements at the 5'-(promoter) and 3'-(E-element) region of the 35S rDNA, including the core element (CE), the upstream element (UE), replication fork barrier In yeast, Cdc14 and Sir2 associate together with the nucleolar protein Net1, forming the "REgulator of Nucleolar silencing and Telophase" (RENT) complex [26][27][28]. Cdc14 and Sir2 interaction interfaces have been mapped to the N-terminal and central parts of full-length Net1, respectively ( Fig 1B) [29][30][31]. By interacting with Cdc14, Net1 inhibits its phosphatase activity and sequesters the protein in the nucleolus. Presumably triggered by specific phosphorylation events Net1 releases Cdc14 in late anaphase. Cdc14 release in turn, leads to dephosphorylation of nuclear and cytoplasmic substrates, which facilitate completion of anaphase, cytokinesis, and progression into G1 [32][33][34][35][36]. Net1 acts likely as a recruiting factor for Sir2, to establish rDNA silencing of Pol II transcription within the intergenic sequence 1 and 2 (IGS1+2, Fig 1A, middle panel, also named NTS1+2, for non-transcribed spacers 1 and 2, in the literature) [27,[37][38][39]. Accordingly, main association sites for the RENT complex have been mapped to the 35S rDNA promoter region in IGS2, and to the replication fork barrier sequence (RFB) in IGS1 at the 3' end of the 35S rRNA gene (Fig 1A) [37]. RENT association with the RFB depends on the fork blocking protein Fob1, presumably through interaction with the Net1 N-terminus [37,40]. Efficient association with the 35S rDNA promoter, instead, requires UAF [41].
A presumably RENT complex-independent function of Net1 in activating Pol I transcription was reported more than 15 years ago [42]. Thus, temperature sensitivity of net1 mutants, in which rRNA production was impaired, could be suppressed by overexpression of the Pol I initiation factor Rrn3. Additionally, recombinant purified Net1 stimulated Pol I transcription in vitro. The molecular basis for these observations, however, remains far from being understood. In this study, we aimed at investigating how Net1 stimulates Pol I transcription. Strikingly, we discovered that the C-terminal 138 amino acids of Net1 harbored the Pol I transcription stimulating function. The C-terminal region (CTR) activated Pol I transcription outside the context of the full-length Net1 protein in vivo and was sufficient to stimulate Pol I transcription in vitro. The identification of the CTR of yeast Net1 as a Pol I transcription activator made it possible to discover similarities of this protein domain with the acidic tail region of the human Pol I transcription factor UBF. Our results are discussed in the light of apparent conservation of the Pol I transcription machineries in distantly related organisms.

Deletion of the C-terminal 138 amino acids of Net1 significantly impairs growth
We established haploid yeast strains, expressing a series of Net1 truncation mutants C-terminally fused to GFP from the endogenous NET1 locus (Fig 2A, cartoon on the right, S1A Fig,  western blot analysis). These yeast strains were subjected to growth analyses on solid medium and in liquid cultures (Fig 2A and 2B; S1 Dataset). A strain expressing full-length Net1 protein fused to GFP grew as well as a NET1 wild-type strain ( Fig 2B and S1B Fig, compare NET1 with net -GFP ), indicating that the GFP tag likely did not interfere with important functions of the protein. C-terminal truncation of the Net1 sequence by only 138 amino acids led to a significant slow-growth phenotype (Fig 2A and 2B, compare net1 GFP (1-1051) with net -GFP ). Additional truncation affecting the integrity of the Sir2-binding domain of Net1 increased the doubling time ( Fig 2B, net1 GFP (1-693), , and (1-341)). A strong growth defect, indistinguishable from that of a NET1 deletion strain, was observed when the C-terminal truncation affected the Cdc14-binding domain (Fig 2A and 2B, compare net1 GFP  with net1Δ). This suggested that the last 138 amino acids of Net1 (hereafter referred to as Cterminal region, CTR) harbor an important growth-supporting function.

A CTR deletion mutant still associates with Cdc14 and Sir2, and localizes to the nucleolus
A mutant Net1 protein lacking the CTR (hereafter referred to as Net1ΔCTR) still contains the regions important for Net1 binding to Cdc14 and Sir2 in the context of the RENT complex ( Fig 1B) [29][30][31]. In good agreement, both FLAG-tagged full-length Net1 (Net1 FLAG ) and Net1ΔCTR (Net1ΔCTR FLAG ) fusion proteins co-precipitated Myc-epitope-tagged Cdc14 and Sir2 (Cdc14 myc , Sir2 myc ) with similar efficiencies from whole-cell extracts (Fig 2C, compare lanes 4 with 5, and 10 with 11; S1C Fig, full membrane). As expected, FLAG antibody-mediated precipitation of Myc-tagged Cdc14, or Sir2, was not observed when whole-cell extracts were prepared from strains that did not express FLAG fusion protein (Fig 2C, lanes 6 and 12). We then examined if the CTR was required for the reported nucleolar localization of Net1 [26][27][28]. Live cell imaging of haploid and diploid yeast strains co-expressing Net1 GFP fusion proteins and the nucleolar marker protein Nop56 fused to mCherry (Nop56 mCherry ) confirmed co-localization of both proteins in the crescent shaped yeast nucleolus (Fig 2D and 2E). Net1ΔCTR GFP showed clear nucleolar localization in a diploid yeast strain expressing a NET1 wild-type allele ( Fig 2F). This strain was not compromised in growth (not shown), indicating that the net1Δctr GFP allele had no dominant negative effect. A haploid strain expressing Net1ΔCTR GFP was impaired in growth (Fig 2A and 2B), and showed abnormal cell morphology (Fig 2G, DIC), and de-localization of Nop56 mCherry over the whole nucleus (Fig 2G, mCherry). However, a subpopulation of Nop56 mCherry preferentially co-localized with Net1ΔCTR GFP within a subnuclear compartment, likely the remainder of the nucleolus (Fig 2G, GFP and merge).

Expression of the CTR in trans rescues growth defects in net1Δctr and net1Δ strains
To analyze if the CTR is an independent functional domain, we tested whether expression of the CTR in trans could rescue the phenotypes observed in a net1Δctr strain. A cassette for constitutive overexpression of the CTR N-terminally fused to GFP ( GFP CTR) under the control of the TEF2 promoter was stably integrated in the LEU2 locus. Co-expression of GFP CTR in trans fully restored wild-type growth in net1Δctr and net1(1-455) strains, and largely suppressed the growth defect of a net1Δ strain (Fig 3A and 3B; S1 Dataset). This indicated that the CTR and the first 455 amino acids of Net1, containing the Cdc14 binding region (but lacking the Sir2 interacting domain) are sufficient to promote normal cell growth. Live cell imaging showed a preferential nucleolar localization of GFP CTR in all strains, although the overexpressed fusion protein also spread all over the cell (

Deletion of the CTR impairs Net1 association with the 35S rDNA promoter
Net1 and other RENT components associate with two rDNA regions, the 35S rDNA promoter region and the RFB at the 3'-end of the 35S rRNA gene [37]. To investigate if deletion of the CTR influenced this binding pattern, we performed Chromatin Endogenous Cleavage (ChEC) [43] and Chromatin Immuno-Precipitation (ChIP) analyses. Haploid yeast strains expressing either wild-type Net1 or Net1ΔCTR in fusion with micrococcal nuclease (MN) and a hemagglutinin (HA) epitope (Net1-MN HA , ΔCTR-MN HA ) were created. For ChEC analyses, crude nuclei isolated from exponentially growing formaldehyde-crosslinked cells were treated with calcium to activate the MN. Cleavage events of the MN HA -fusion proteins were mapped to the indicated genomic regions by Southern blot analysis using the indirect end-labeling technique [44]. Net1-MN HA fusion protein mediated cleavages were observed both at the 35S rDNA promoter and at the RFB in good correlation with earlier results (Fig 4A, lanes 1-4 and 9-12) [41]. Net1ΔCTR-MN HA -mediated cuts were strongly reduced at the 35S rDNA promoter region, whereas robust cleavages were still observed at the RFB (Fig 4A, lanes 5-8 and 13-16). In good agreement, Net1-MN HA but not Net1ΔCTR-MN HA co-precipitated 35S rDNA promoter fragments from cellular extracts in ChIP experiments (Fig 4B, compare ChIP of fragment 2, cartoon at the bottom for location of ChIP fragments in yeast rDNA). In contrast, co-precipitation of RFB fragments with Net1ΔCTR-MN HA was unaltered or even slightly , and y3290 and y3298 (B)) were included as controls. C, D) The diploid yeast strain W15407 was sporulated, yielding haploid progenies carrying either a NET1 wild-type allele (C) or a net1Δctr allele (D), and expressing Nop56 mCherry and GFP CTR. These haploid strains were subjected to live cell fluorescence microscopy as described in the legend to Fig 2D. https://doi.org/10.1371/journal.pgen.1008006.g003 enhanced, when compared to ChIP with Net1-MN HA (Fig 4B, compare ChIP of fragments 5, 6). As expected, no significant co-precipitation of rDNA fragments bearing the 5S rDNA, 18S rDNA, and 25S rDNA regions with the fusion proteins was observed (Fig 4B, ChIP of fragments 1, 3, 4). Thus, it can be concluded that the CTR is required for interaction of Net1 with the 35S rDNA promoter but does not affect Net1 association with the RFB, which is mediated by Fob1 [37]. In good correlation, Fob1 association with the RFB was unaltered or even slightly increased in net1Δctr strains when compared with NET1 strains (  Haploid yeast strains expressing the indicated proteins C-terminally fused to MN HA were subjected to ChEC, or ChIP experiments. A, C) Crude nuclei suspensions were subjected to ChEC in the presence of calcium for the indicated times (min ChEC), DNA was isolated, cut with XcmI, separated in a 1% agarose gel and subjected to Southern blot analysis by indirect end-labeling using the radiolabeled probes rDNp, IGS2. An autoradiograph of the membrane is shown. Cartoons at the sites depict a map of the detected genomic fragments including probe hybridization sites and important cis elements described in the legend to Fig 1A. Arrows at the top of the cartoons point to the uncut full-length XcmI fragments. An asterisk marks the site of a DNA double-strand break occurring at the 3' end of the RFB in the absence of ChEC. B, D) fragmented chromatin was subjected to ChIP with the indicated MN HA tagged fusion proteins using anti-HA (3F10) antibody. DNA coprecipitating with the tagged proteins and from input fractions was isolated and analyzed by quantitative PCR with different primer pairs amplifying distinct regions (1-6) within the rDNA locus indicated in the cartoon at the bottom. Bar graphs depict the efficiency of co-precipitation of the fragments as percentages of the input DNA (% ChIP). Mean values and standard deviation errors were derived from three independent ChIP experiments analyzed each in triplicate qPCR reactions (n = 9). A, B) ChEC and ChIP analyses with strains (y3041, y3042, y3043, y3044) expressing either Net1-MN HA or Net1ΔCTR-MN HA (ΔCTR-MN HA ). C, D) ChEC and ChIP analyses with strains (y3058, y3068, y3250), carrying a NET1 or a net1Δctr allele, and expressing Cdc14-MN HA , in the absence or presence of a chromosomally integrated FLAG CTR expression cassette.

35S rDNA promoter association of RENT complex components is impaired in the net1Δctr strain
Cdc14 and Sir2 have rDNA association patterns very similar to that of Net1 [37,41,45]. Accordingly, Cdc14-MN HA mediated ChEC in a NET1 wildtype strain led to cuts within the 35S rDNA promoter region and at the RFB, in good correlation with enrichment of these regions in ChIP experiments (

Expression of the CTR in trans does not rescue 35S rDNA promoter interaction of RENT complex components in the net1Δctr strain
The impaired association of RENT complex components with the 35S rDNA promoter could be causative for the growth defect observed in the net1Δctr strain. Ectopic expression of FLAG CTR in net1Δctr strains restored wild-type growth (Fig 3A and 3B). Thus, it was tested if the interaction of Cdc14 with the 35S rDNA promoter could be re-established in this condition. ChEC and ChIP experiments were carried out in a net1Δctr strain expressing Cdc14-MN HA from the endogenous locus and the FLAG CTR from a chromosomally integrated cassette. Expression of FLAG CTR resulted neither in significant Cdc14-MN HA -mediated cleavage at the 35S rDNA promoter in ChEC experiments, nor did it lead to coprecipitation of 35S rDNA promoter fragments in ChIP (

Pol I association with 35S rRNA genes requires the CTR
The above experiments suggested, that the CTR plays a RENT complex-independent role supporting cellular growth, presumably by interacting with the 35S rDNA promoter. We therefore tested if the CTR harbors the Pol I transcription stimulating function of Net1. Haploid yeast strains carrying a NET1 or a net1Δctr allele and co-expressing MN HA fusion proteins of the Pol I subunits Rpa43 or Rpa190 were subjected to ChEC and ChIP experiments. ChEC analyses revealed that Rpa43-MN HA and Rpa190-MN HA cleavage events at the 35S rDNA promoter region and at the 35S rDNA transcription termination site upstream of the RFB were reduced in net1Δctr strains, when compared to cleavages in NET1 strains ( Table, doubling times).

The CTR interacts with many chromosomal loci
Since the CTR is required to recruit full-length Net1 to the 35S rDNA promoter (Fig 4A), we further investigated if the isolated CTR expressed in trans may interact with this chromosomal region. ChEC experiments were performed in NET1 or net1Δctr strains, expressing the CTR with an N-terminal FLAG tag and C-terminally fused to MN HA ( FLAG CTR-MN HA ), under the control of the TEF2 promoter from a cassette integrated in the LEU2 locus. As observed for GFP CTR and FLAG CTR, expression of the FLAG CTR-MN HA fusion protein suppressed the slow-growth phenotype of net1Δctr strains (S3 Table). Similar FLAG CTR-MN HA -mediated cleavage events within the 35S rDNA promoter region were observed in both NET1 and net1Δctr strains (Fig 6A, upper panels, lanes 1-4, and 5-8). FLAG CTR-MN HA -mediated cuts were also observed at the RFB and at further sites within IGS1 (Fig 6A, lower panels, lanes 1-4, and 5-8). Additionally, FLAG CTR-MN HA -dependent cleavages were observed at accessible  Thus, upon overexpression of the CTR, the protein is available for DNA interactions within the entire nucleus, including the 35S rDNA promoter.

Expression of a Rpa190-CTR fusion protein suppresses the growth defect of a net1Δctr strain
The above data are compatible with a model in which the CTR assists Pol I recruitment at the 35S rDNA promoter. To further test this hypothesis experimentally, we fused the FLAG CTR to the C-terminus of Rpa190 and the CF component Rrn7 in net1Δctr strains. Whereas expression of Rrn7-FLAG CTR could not significantly rescue the growth phenotype of net1Δctr strains, Agarose gels were stained with SYBR safe before transfer to the blotting membrane (upper panels), sizes of selected DNA fragments in the DNA marker (M) are given on the left. Lower panels, the membrane was hybridized with a radioactively labeled probe hybridizing with a XcmI fragment containing the NUP57/RPS23A loci as indicated in the cartoon on the left. C, D) Haploid yeast strains (y3673; y3674; y3746; y3747; y3748; y3749; y3756; y3757; y3758; y3759, y3290, y3298) carrying a NET1 or a net1Δctr allele and expressing CTR fusion proteins of Rpa190, or Rrn7 or the wild-type proteins, were subjected to growth analyses in liquid cultures (C), and on solid medium (D) as described in the legend to Fig 2A and  expression of Rpa190-FLAG CTR fully suppressed slow growth (Fig 6C and 6D; S1 Dataset). The expression of CTR fusion proteins had no impact on growth in NET1 strains. This observation is in accordance with a Pol I-specific function of the CTR, but also indicates that simple tethering of the CTR to the 35S rDNA promoter may not be sufficient to obtain full stimulation of Pol I transcription.

Pol I associates with phosphorylated CTR
Previous work showed co-purification of Net1 with Pol I subunits [42,46]. Here, it was tested if the CTR of Net1 could associate with Pol I. Immuno-precipitation experiments using wholecell extracts from a strain expressing Rpa43 myc and FLAG CTR with an anti-Myc antibody showed enrichment of the Myc-tagged bait protein and co-precipitation of a minor sub-population of FLAG CTR in western blot analysis (Fig 7A, lanes 1 and 3). As expected, FLAG CTR was not precipitated from whole-cell extracts of strains expressing the untagged Rpa43 protein ( Fig  7A, lanes 2 and 4). FLAG CTR in whole-cell extracts migrated as a diffuse band around 30 kDa FLAG CTR in WCE of the diploid yeast strain W13490 was bound to anti-FLAG M2 antibody agarose and treated with increasing amounts (grey triangle) of λ-protein phosphatase (λ-PP). Protein samples were analyzed by western blot as described in the legend to Fig 2C. C) The CTR is differentially phosphorylated in vivo. Haploid yeast strain y3739 expressing FLAG CTR was grown to exponential (exp), or to stationary phase (stat). WCE was prepared and subjected to western blot analysis with anti-FLAG antibody (αFLAG) as described in the legend to Fig 2C. Ponceau S stain of the membrane, and an image visualizing chemiluminescence signals upon immunodetection (αFLAG) are shown. D) Recombinant CTR expressed in insect cells is hyperphosphorylated. Net1 CTR TAP expressed in insect cells was bound to IgG-magnetic beads. Net1 CTR TAP bound to the beads was either mock treated (-) or incubated with λ-PP (+). CTR CBP fusion proteins were released by TEV protease cleavage and analyzed by SDS-PAGE and Coomassie staining. A photograph of the gel is shown. The size of molecular weight markers in kDa is indicated on the left. Positions of the tagged proteins as well as TEV protease are indicated on the right. E) CTR-dependent stimulation of Pol I transcription in vitro is decreased after λ-PP treatment. In vitro transcription reactions were carried out using Pol I purified from yeast, recombinant CF and Rrn3 purified from E. coli in the absence (-) or presence of increasing amounts (grey triangle) of CTR CBP , either mock treated or incubated with λ-PP as described in (D). Radiolabeled RNA was isolated and separated by 6% acrylamide-urea gel electrophoresis. An autoradiograph is shown. The position of the specific transcript is indicated on the right. in SDS-PAGE, despite an estimated molecular weight of 18 kDa. Net1 is known to be phosphorylated and several phosphorylation sites have been identified within the CTR in proteomic studies (Fig 8A, asterisks mark described phosphorylation sites) [32,[47][48][49][50]. To investigate if the aberrant migration behavior of the CTR could be explained by phosphorylation, FLAG CTR was enriched from whole-cell extracts using an anti-FLAG antibody matrix. The immobilized protein was incubated in the absence or presence of different concentrations of lambda protein phosphatase (λ PP). Western blot analysis revealed that the incubation with λ PP substantially increased the electrophoretic mobility of FLAG CTR (Fig 7B, lanes 3-5; S4B Fig, full membrane) suggesting that the CTR is phosphorylated in vivo.

Changes in post-translational covalent modifications of the CTR correlate with alterations in Pol I transcription of 35S rRNA genes in vivo
To investigate if the post-translational modification state of the CTR was subject to changes in vivo, whole-cell extracts were prepared from exponentially growing and stationary cells expressing FLAG CTR. Proteins were separated by SDS-PAGE and analyzed in a western blot. FLAG CTR in extracts from stationary cells had a higher mobility in SDS-PAGE than FLAG CTR in extracts from exponentially growing cells, suggesting alterations in the modification state (Fig 7C, αFLAG, compare "exp" with "stat"; S4C Fig, full membrane). Ponceau-red staining of the immunoblot membrane indicated that the overall mobility of cellular proteins was not altered in the different extracts (Fig 7C, Ponceau; S4C Fig). Ribosome biogenesis and Pol I transcription are downregulated when yeast cells grow to stationary phase [51,52]. Downregulation of Pol I transcription correlates with a decreased psoralen accessibility of the 35S rRNA gene region transcribed by Pol I [53][54][55]. As a control for downregulation of Pol I transcription in the above experiment, we subjected cells from the same cultures used for protein extraction to psoralen crosslinking analysis. In agreement with the expected alterations in Pol I transcription, psoralen accessibility of 35S rRNA genes was high in exponentially growing cells and negligible in stationary phase cells (S4D Fig, compare lanes 4 and 5).

λ protein phosphatase treatment decreases CTR-dependent stimulation of promoter-dependent Pol I transcription in vitro
Recombinant full-length Net1 has been shown to stimulate promoter-dependent Pol I transcription in vitro [42]. We tested if the CTR could be sufficient to promote Pol I stimulation in a reconstituted system. The CTR was expressed in insect cells with a C-terminal tandem affinity purification (TAP) tag (CTR TAP ) bound to IgG-sepharose and released as a calmodulin binding peptide (CBP) fusion protein (CTR CBP ) by cleavage with tobacco etch virus (TEV) protease (S4E Fig) [56]. As observed for the CTR expressed in exponentially growing yeast (Fig 7A-7C), purified recombinant CTR CBP with a predicted size of 20 kDa migrated with an apparent molecular weight of 32 kDa in SDS-PAGE (Fig 7D; S4F Fig, lane1; S4E Fig, lane 3). λ PP treatment prior to TEV-mediated release from IgG-Sepharose significantly increased the mobility of the recombinant protein in SDS-PAGE when compared with the mobility of the mock-treated CTR CBP , indicating that the protein was phosphorylated (Fig 7D compare lane 1 with lane 2; S4F Fig, full  gel). λ PP-treated and mock-treated CTR CBP were tested in promoter-dependent Pol I transcription in a minimal in vitro system consisting of Pol I purified from yeast and of recombinant Rrn3 and CF purified from E. coli [57]. The mock treated CTR CBP could stimulate promoterdependent Pol I transcription in a concentration-dependent manner up to 6-fold, whereas the λ PP treated CTR CBP still led to a 2 to 3-fold stimulation (Fig 7E, compare lanes 1 and 2, with  lanes 3-6 and lanes 7-10). Additionally, recombinant full-length Net1 TAP and Net1ΔCTR TAP fusion proteins were purified from insect cells (S4G Fig) and tested in parallel with another preparation of CTR CBP protein with and without λ PP treatment. The recombinant Net1 CBP was the strongest activator of Pol I transcription, whereas Net1ΔCTR CBP still stimulated Pol I transcription to a similar extent as the purified CTR CBP (S4H Fig). Thus, it is possible that Net1ΔCTR still partially supported Pol I transcription in vivo. In agreement with such hypothesis, residual Net1ΔCTR-MN HA -mediated cleavage at the 35S rDNA promoter could still be observed upon ChEC (Fig 4A, lanes 5-8; S3D Fig, upper panel, lanes 5-8). In contrast, ChEC analyses performed with cells from a Net1(1-341)-MN HA expressing strain did not lead to significant cuts within the 35S rDNA promoter (S3D Fig, upper panel, lanes 9-12). Net1(1-341)-MN HA mediated cleavage at the RFB region, instead, was comparable-although not identical-to cleavage mediated by Net1-MN HA or by Net1ΔCTR-MN HA (S3D Fig, lower panels).

The acidic region of human UBF1 can in part functionally substitute for the CTR in the context of the Net1 protein
Whereas the full-length Net1 is not conserved in higher eukaryotes, local sequence alignment using the tool "matcher" of the EMBOSS platform revealed 25% identical and 43.5% similar residues within the acidic tail region (AR) of human UBF1 Pol I transcription factor (Fig 8A). An acidic serine/aspartate-rich sequence within the CTR showed the highest similarity with the UBF1 AR (Fig 8A, underlined amino acids). Assuming, that the CTR is hyperphosphorylated within this sequence, the similarity between the two regions might be even greater. To test if the UBF1 AR could functionally substitute for the Net1 CTR, yeast strains were created expressing different derivatives of Net1 in fusion with MN HA (Fig 8B). Full-length Net1, Net1ΔCTR, or Net1 variants in which the CTR was replaced by the AR of human UBF1, an Nterminally extended AR (eAR), or a triplicated synthetic minimal VP16 activation domain (VP), respectively were analyzed. The VP16 activation domain triggers high-level Pol IIdependent transcription when recruited to gene promoters by a specific DNA binding domain in higher eukaryotes [58]. The expression of the respective MN HA -fusion proteins in the above strains was verified by western blot analysis (Fig 8C; S5A Fig, full membrane). Six independently obtained clones of each of these strains were subjected to growth analyses together with clones of a strain expressing an untagged Net1 protein as a reference control (Fig 8D; S1 Dataset). Whereas all the strains expressing Net1 variants lacking the CTR grew slower than the Net1-MN HA or Net1-expressing strains, Net1ΔCTR-AR-MN HA and Net1ΔCTR-eAR-MN HAexpressing strains grew slightly but statistically significantly faster than Net1ΔCTR-MN HAexpressing strains. This indicated that the Net1-CTR and the UBF1-AR regions share a conserved function in promoting growth. In contrast, expression of Net1ΔCTR-VP-MN HA did not significantly rescue the net1Δctr growth phenotype in those experiments (Fig 8D).

Discussion
In this study we could attribute the Pol I stimulating function of the multi-functional yeast Net1 protein to its C-terminal 138 amino acids (CTR). In good accordance with the presumed role of Pol I transcription in promoting cell proliferation, the CTR supported wild-type growth (Figs 2A, 2B, 3A and 3B; S1 Dataset). The CTR was further required for normal cellular morphology and localization of the nucleolar protein Nop56 (Figs 2G and 3D, S2A and S2B Fig). It is likely that the observed nucleolar de-localization of Nop56 in net1Δctr strains can be explained by impaired Pol I transcription. This would be in agreement with earlier observations showing that rRNA gene transcription by Pol I is an important determinant for proper localization of nucleolar components [59][60][61].
The nucleolar localization of Net1 correlates with its reported interaction with two different regions within the rDNA [37]. The CTR was dispensable for nucleolar localization of Net1 and for recruitment of Net1 and Cdc14 to the RFB element at the 3'-end of the 35S rRNA gene (Figs  2G and 4; S3D Fig). The latter observations are in good agreement with the observation that only the N-terminal 341 amino acids of Net1 are needed for its interaction with Fob1 [62], which in turn is needed for Net1 interaction with the RFB [37]. In contrast, the CTR was required for Net1 and Cdc14 association with the 35S rDNA promoter (Fig 4; S3D Fig). Secondary structure prediction did not provide evidence for the presence of a DNA binding domain within the CTR, but rather suggested that this protein region might be intrinsically disordered (S5B Fig). Thus, it is more likely that the CTR interacts with the promoter region in context of the DNA-bound protein (RNA) scaffold. Interestingly, interaction of Net1 with the 35S rDNA promoter was impaired in a uaf30Δ strain in which Pol I PIC formation is strongly affected [41], indicating that Net1 may interact with PIC components. Candidate interaction partners could be the UAF subunits Rrn5 and Rrn10, since they were found in a Cdc14 interactome study in which the entire RENT complex-including Net1-was purified [63]. Additionally, co-purification of many Pol I subunits with full-length Net1 was reported [37,42,64], and here, co-purification of the CTR with the Pol I subunit Rpa43 was shown (Fig 7A). A Pol I-related function of the CTR was further supported by the observation that the Pol I subunit Rpa190 in fusion with the CTR could suppress slow growth of net1Δctr strains (Fig 6C and 6D). Net1 interaction with the 5' end of the 35S rDNA is not restricted to the promoter region but extends into the external transcribed spacer 1 as revealed in ChEC and ChIP experiments (Fig 4A; S3D Fig) [37,41]. Such a trailing into the Pol I transcribed 35S rDNA resembles the association observed for the Pol I transcription factor Rrn3 [65]. Thus, the CTR might interact with promoter-bound factors and assist efficient PIC assembly, and possibly additional downstream events of the Pol I transcription process.
The amino acid sequence analysis of the CTR revealed similarity with the AR of the human Pol I transcription factor UBF1 (Fig 8A). However, it should be noted that the CTR amino acid sequence is of low complexity, mainly composed out of aspartate and serine residues. Additionally, we used the HHPred module from the free HH-suite software package [66,67] to detect homologous domains in other proteins which identified a very similar (small) set of RNA-binding proteins for both CTR and AR (not shown). In our experiments, the AR of human UBF1 could at least partially substitute for the growth-supporting function of the CTR in the context of Net1 (Fig 8D; S1 Dataset). However, the observed effects on growth were small. Accordingly, Net1-(e)AR-MN HA fusion proteins could not significantly restore cleavage at the 35S rDNA promoter in ChEC experiments (S5C Fig). Several studies have provided evidence that the carboxyterminus of UBF1 is important to support efficient Pol I transcription, probably by stabilizing the UBF1-SL1 complex at the rDNA promoter [68][69][70]. Interestingly, the AR of UBF1 is also required for proper nucleolar localization of UBF1 and preferentially localizes to the nucleolus when expressed in trans [71,72], alike observations with the CTR of Net1 (Fig 3C and 3D; S2  Fig). Phosphorylation of specific residues within the acidic tail have been suggested to increase UBF-SL1 interaction [38,39,[73][74][75], and UBF phosphorylation correlates well with active rRNA gene transcription in vivo [70,73,[76][77][78][79][80][81]. Additionally, de-phosphorylation of UBF1 reduces its potential to stimulate promoter-dependent Pol I transcription in vitro [70,80]. Interestingly, a growth-dependent alteration in the post translational modification state could also be observed for the CTR, correlating with changes in Pol I transcription (Fig 7C; S4C and S4D Fig). Furthermore, de-phosphorylation decreased CTR-mediated stimulation of Pol I transcription in vitro (Fig 7E; S4H Fig). As a next step, it will be interesting to identify enzymes and pathways regulating the differential post-translational modification of the CTR. Two Cdc28 (Cdk1 in human) target sites reside within the CTR [32,48], and many residues within the serine-rich region are predicted to be potential substrates for casein kinase 2 (CK2; Fig 8A, predicted CK2 phosphorylation sites in bold [82,83]). Noteworthy, CK2 activity might have a function in the regulation of Pol I transcription [38,84,85]. The possibility that the CTR of Net1 functions in part similarly to the AR in UBF1 could point to further conserved mechanisms in Pol I transcription. As outlined in the introduction, the yeast HMG-box protein Hmo1 might also share functions with HMGboxes 1 and 2 of UBF1 [17]. This could indicate that similar functional domains act in yeast and higher eukaryotes to support Pol I transcription but are organized in different polypeptides.
Recent advances in structure determination of the Pol I transcription machinery have yielded great insights in the architecture of individual components as well as in the assembly of a minimal PIC (reviewed in [86]). In the future, analyses using defined in vitro systems will further deduce structure-function relationships important to eventually understand the mechanism driving Pol I transcription. For faithful reconstitution of this process, it will be necessary to identify all of the participating factors in vivo. In this study, we identified another, likely conserved, component of the yeast Pol I transcription machinery.

Oligonucleotides, plasmids and yeast strains
Unless noted otherwise, standard techniques were used for cloning of plasmids, and transformation of yeast cells [87][88][89]. Information about oligonucleotides, plasmids, and yeast strains used in this study can be found in S1-S3 Tables, including details on plasmid and strain construction. Plasmid sequences are available upon request. For the individual experiments in this study, yeast cells were grown at the indicated temperatures in either YPD (2% w/v) peptone, 1% (w/v) yeast extract, 2% glucose), YPAD (YPD including 100mg/l adenine), or XYD medium (YPD including 100mg/l adenine, 200mg/l tryptophan, and 10mM KH 2 PO 4 ).

Growth analyses
Spot assays. Cultures were grown at 25˚C in XYD media overnight (NET1) or for 3 days (net1-mutants), respectively. Cells were collected by centrifugation and resuspended in sterile water to a final OD 600 of 1. 10-fold serial dilutions of cell suspensions were spotted onto XYD agar plates and incubated at 25˚C for several days. For experiments shown in Fig 6D cultures were grown for two days at 30˚C in YPAD and spotted on YPAD plates. Photographs were taken at different times during the incubation to document growth.
Growth in liquid culture. Cultures were grown in YP(A)D at 30˚C overnight. Cells were sedimented, washed with sterile water and inoculated into 200μl YPD in a sterile 96 well culture plate to a starting OD 600 of 0.02. Incubation was carried out at 30˚C under occasional shaking using a TECAN infinite 500 or 200Pro plate reader system (TECAN). The OD 600/613 was determined every 15 min. Doubling time (τ) was calculated for the exponential growth phase by the formula τ = (t 2 -t 1 )/(log(2)/log(OD t2 /OD t1 ) for more than 17 different time intervals. Mean doubling times and standard deviation errors for most of the strains used in this study can be found in S3 Table. All raw data of the TECAN measurements including growth diagrams shown in Figs 2B, 3B, 6D and 8D can be found in the S1 Dataset.
Statistics. Statistical analyses were performed with GraphPad Prism 5.04. Differences between groups were tested using the nonparametric two-sided Mann-Whitney U (MWU) test. P-values <0.05 were considered statistically significant.

Live cell fluorescence microscopy
Exponentially growing cells were harvested by centrifugation, transferred to glass slides and covered with an agarose slice (supplied with yeast nitrogen base, amino acids, and 2% glucose). Cells were observed at 20-22˚C on an Axio Observer Z.1 confocal spinning disk (CSU-X1) microscope (Yokogawa; Carl Zeiss) using a Plan Apochromat 63×/1.40 oil differential interference contrast (DIC) M27 lens. Ten Z-stack images with an optical section spacing of 0.5 μm were acquired with a shutter speed of 200 ms using an AxioCam MRm camera (Carl Zeiss). Microscopy data was processed using ImageJ (National Institutes of Health, Bethesda, MD). Z-stacks of the region of interest (fluorescence channels) were projected using maximumintensity projection. A single plane (DIC channel) was duplicated to visualize cell morphology.

Chromatin Endogenous Cleavage (ChEC), and Chromatin Immuno-Precipitation (ChIP)
For ChEC and ChIP experiments yeast strains expressing different genes as fusion proteins with a C-terminal MN followed by a triple HA-tag from their endogenous genomic location were grown in YP(A)D at 30˚C to a final OD 600 of 0.5. ChEC analyses were performed as previously described [15]. All DNA samples were digested with XcmI prior to agarose gel electrophoresis and Southern blot analysis (see below). ChIP experiments were performed as described and analyzed by quantitative PCR [41]. Data was collected with a Rotor-Gene Q system (Qiagen). The cycle threshold values and amplification efficiency (E) for a defined primer pair for all input and ChIP samples of an experimental dataset were determined with the Rotor-Gene 6000 software (Qiagen) using the comparative quantification method. The cycle threshold (CT) value correlates with the Rotor-Gene software "take-off" value (tov). For each primer pair the efficiency (E) is calculated as an average from the fluorescence measurements for each individual sample. The percentage of co-precipitated DNA was determined by the formula %ChIP = ((1+E) tov(ChIP) /(1+E) tov(input) ) � 100. Average and standard deviation errors for ChIP are derived from three independent ChIP experiments, each analyzed in triplicate qPCRs. The original qPCR data, including CT, "take off" and efficiency values, melting curve analysis, and mathematical operations are added as S2 Dataset.

Gel electrophoresis and blot analysis
SDS-PAGE and Western blot analysis was performed according to standard procedures [90,91]. Gels were either stained with Coomassie blue for detection of proteins or transferred to nitro cellulose (GE Healthcare) or PVDF membranes (Immobilon P, Roth) for subsequent immuno-detection. Transfer to membranes was verified by staining with Ponceau S. S4 Table  contains a complete list of antibodies used for detection. Secondary antibodies coupled to IRDye 800 were detected with an LI-COR Odyssey Infrared Imaging System (LI-COR Bioscience). Secondary antibodies coupled to horse reddish peroxidase (HRPO) were visualized using BM Chemiluminescence Western Blotting Substrate (POD, Roche), and a LAS-3000 Chemiluminescence Imager.
Agarose gel electrophoresis and Southern blot analysis were performed as described [88], with the exception that transfer of nucleic acids onto nylon membranes (Positive Membrane, Qbiogen) by capillary transfer was performed with 1 M ammonium acetate [92]. S5 Table contains a list of templates used for the synthesis of radiolabeled hybridization probes generated using the RadPrime DNA labeling system (GE Healthcare).

Protein extraction, immuno-precipitation (IP) and λ-phosphatase treatment
For preparation of native protein extracts yeast cells were grown overnight in XYD media at 25˚C to an OD 600 of 0.8. For preparation of yeast whole-cell extracts (WCE), cell equivalents corresponding to 10-20 OD 600 were harvested by centrifugation (805 rcf, 2 min) and washed once with ice-cold water. Lysis buffer (150 mM NaCl (for IPs with Net1 FLAG -fragments) or 350 mM KCl (IPs with Rpa43 myc ), 50 mM Tris-HCl pH 7.5, 50 mM NaF, 5 mM EDTA, 0.1% IGEPAL CA-630, 60 mM β-glycerol phosphate) was added to yield a final volume of 150 μl cell suspension. After addition of 200μl of glass beads cells were lysed by shaking the mixture for 5 min at 4˚C in a mixer mill (Retsch). After removal of cell debris by two consecutive centrifugation steps of 10 min and 15 min (16000 rcf, 4˚C), equal volumes of supernatant and 2× SDSsample buffer were mixed and incubated for 10 min at 100˚C.
For immuno-precipitation of Myc-epitope tagged RNA polymerase I subunits, WCEs (adjusted to a total volume of 450μl) were incubated with anti-Myc antibodies (clone 9E10) for 2h at 4˚C. Thereafter, 50 μl of protein A-agarose slurry (Santa Cruz Biotechnology Inc.) were added, and the mixtures were incubated on a rotator for 2 h at 4˚C. For immuno-precipitation of FLAG-tagged Net1-fragments, WCEs (adjusted to a total volume of 450 μl) were incubated with 45μl of α-FLAG-beads (Sigma-Aldrich) on a rotator for 3 h at 4˚C. Beads were collected by centrifugation and washed three times with 500 μl lysis buffer. After the last washing step, beads were resuspended in 1x SDS sample buffer and incubated for 10 min at 100˚C.
After immuno-precipitation of FLAG-tagged CTR as described above, beads were collected by centrifugation and washed twice with 1ml of lysis buffer (without β-glycerol phosphate) and once with λ-phosphatase buffer (50 mM HEPES-KOH pH 7.5, 5 mM DTT, 2 mM MnCl 2 ). Beads were split in different samples and treated in the absence (control), or presence of 200, 400, and 2000 units of λ-phosphatase (New England Biolabs) in a total volume of 50μl λ-phosphatase buffer, and samples were incubated for 1 h at 37˚C. Beads were collected by centrifugation and washed twice with 1 ml λ-phosphatase buffer. After the last washing step, beads were resuspended in 1x SDS-sample buffer and incubated for 10 min at 100˚C.
All samples were analyzed by western blot as described above. The data presented in Figs 2C, 7A and 7B are representative for at least three independent experiments.

Yeast cell growth to stationary phase with subsequent protein and trimethyl psoralen crosslinking analysis
Single colonies of yeast strains were used to inoculate 3 ml of YPAD which were incubated for 2 days at 30˚C under shaking. This culture was used to inoculate YPAD media to an OD 600 not higher than 0.1. Samples for denaturing protein extraction or trimethyl psoralen crosslinking analyses were withdrawn when the culture reached an OD600 of 0.5 (exponential phase) or 144 hours after this time point (stationary phase). For protein analysis cell equivalents of 0.5-3 OD 600 , were subjected to protein extraction as described elsewhere [93]. For trimethylpsoralen crosslinking analysis, cell equivalents of 25 OD 600 were treated as previously described [15]. All DNA samples were digested with EcoRI prior to agarose gel electrophoresis and Southern blot analysis (see above).

Purification of recombinant proteins and λ-phosphatase treatment
Purification of 6x histidine-tagged Rrn3, or 6x histidine-tagged core factor from E. coli, as well as TAP-tagged RNA Pol I from yeast have been described previously [57]. Recombinant expression of different Net1-TAP fusion proteins in baculovirus infected S. frugipedia SF21 cells was achieved according to a protocol described in [94][95][96]. For TAP-tag mediated purification, 50x10 6 cells from a large scale infection were suspended in 30 ml TAP-lysis buffer (50 mM HEPES-KOH pH 7.5, 2.5 mM Mg-acetate, 0.001% (w/v) Tween, 0.2 M KCl, 5 mM 2-mercaptoethanol, 1 mM PMSF and 2 mM benzamidine) and whole-cell lysates were prepared as previously described [96]. To purify TAP tagged proteins from 30ml of extract were added to 75mg of magnetic beads (BcMag, Bioclone Inc.) coupled to rabbit IgGs (Sigma) prepared as described [97]. Further purification followed the protocol for TAP-tagged RNA Pol I from yeast [57].
Dephosphorylation of recombinant CTR-TAP fusion protein was performed similarly to the dephosphorylation of the FLAG-tagged CTR purified from yeast (see above). In brief, after binding TAP-tagged CTR to IgG-coated magnetic beads (see above), the affinity matrix was washed 3x with 1 ml TAP-lysis buffer, and 3x with 1 ml TAP-lysis buffer lacking PMSF and benzamidine. The magnetic beads were equilibrated with 1 ml λ-phosphatase buffer and split into two aliquots. Beads in each aliquot were suspended in a total volume of 50 μl λ-phosphatase buffer in the absence (mock control) or presence of 2000 u λ-phosphatase and incubated in a thermomixer (Eppendorf) for 1h at 30˚C under shaking. The supernatant was discarded, and the beads were washed 3x with 1 ml TAP-lysis buffer lacking PMSF and benzamidine. Proteins bound to the beads were eluted by TEV cleavage as previously described [57]. This experiment has been repeated three times and a representative experiment is shown in Fig 7D.

Promoter-dependent RNA polymerase I in vitro transcription
In vitro transcription analysis was performed essentially as previously described [57], with minor changes. To allow formation of the initiation-competent Rrn3-Pol I complex, 5 nM RNA Pol I and 70 nM Rrn3 were incubated at 4˚C for 60 min. The reaction was complemented with 10 nM CF, and 5 nM template DNA containing the 35S rDNA promoter. Additionally, potassium acetate was added to yield a constant salt concentration of 200 mM in the final transcription reaction. In some reactions affinity purified Net1 derivatives were included (5-20 nM for full-length Net1, and Net1ΔCTR, and 20-100 nM for CTR or dephosphorylated CTR). The reaction was eventually adjusted to 25μl containing 20 mM HEPES-KOH pH 7.8, 10 mM MgCl 2 , 5 mM EGTA, 0.05 mM EDTA, 2.5 mM DTT, 0.2 mM ATP, 0.2 mM UTP, 0.2 mM GTP, 0.01 mM CTP including 1.25 μM α-32 P-CTP. Samples were incubated for 30 min at 24˚C, before the reaction was stopped and processed as reported [57]. The data presented in Fig 7E and S4H Fig are representative for at least three independent experiments.   strains and has a preferential nucleolar localization in net1Δ strains. A,B) The diploid yeast strain W15406 was sporulated, yielding haploid progenies carrying a net1(1-455) allele, and expressing Nop56 mCherry in the absence (A) and presence of GFP CTR (B). Haploid strains were subjected to live cell fluorescence microscopy as described in the Legend to Fig 2D. C) The diploid yeast strain W12509 was sporulated, yielding haploid progenies carrying a net1Δ allele, and expressing GFP CTR. The haploid strain was subjected to live cell fluorescence microscopy as described in the Legend to Fig 2D. D,E) FLAG CTR expression suppresses the growth defect of net1Δctr strains and does not alter expression levels of Net1ΔCTR FLAG Haploid yeast strains (y3058, y3068, y3250), carrying a NET1 or a net1Δctr allele, and expressing Cdc14-MN HA in the absence or presence of a chromosomally integrated expression cassette for FLAG CTR were subjected to growth and western blot analyses. D) Growth analyses in liquid culture were performed as described in the legend to Fig 2B. E) WCEs were prepared and subjected to western blot analysis as described in the legend to Fig 2C, Fig 6A-6C (see legend to these figures for more information). D) Haploid yeast strain y3739 expressing FLAG CTR was cultured in YPD. Different samples were withdrawn at exponential (exp) and stationary phase (stat), and either used for protein extraction (Fig 6C; S4C Fig) or treated with formaldehyde for psoralen crosslinking analysis. Crude nuclear extracts were prepared from formaldehyde treated cells, which were subjected to the psoralen crosslinking procedure. DNA was isolated, digested with EcoRI, separated by native agarose gel and subjected to Southern blot analyses with probe "3.5kb rDNA". Positions of rDNA fragments derived from Pol I transcribed "open" 35S rRNA genes and nucleosomal "closed" 35S rRNA genes are depicted on the right. An autoradiography of the Southern blot membrane is shown. Lanes 1 and 2 show the analysis of a control experiment with strain y3725, which did not express FLAG CTR. Lane C included genomic DNA digested with EcoRI, isolated from a strain which was not subjected to psoralen crosslinking but mock treated in parallel with the other strains. The sizes of two DNA fragments in kb spanning either parts of the 18S or the 25S rDNA transcribed by Pol I and visualized by probe "3.5kb rDNA" are depicted on the left. E-H) Recombinant Net1, Net1DCTR and CTR purified from insect cells stimulate promoterdependent Pol I transcription in vitro, and CTR-mediated stimulation of in vitro transcription is reduced upon λ-protein phosphatase treatment E,G) Purification of recombinant Net1 TAP variants from bacmid infected Sf9 cells was performed as described in the legend to Fig 6D. Proteins contained in samples of the whole-cell extract (WCE, 0.06% of total), the flow through (FT, 0.06%), elution 1 and 2 after cleavage with TEV protease (E1, E2, 20%), and beads after elution (B, 20%) were analyzed by SDS PAGE and Coomassie blue staining. Photographs of the gels are shown. Molecular weights and positions of marker proteins are indicated on the left. F) Photograph of the entire gel of the experiment described in the legend to Fig 6D. H) In vitro transcription reactions were carried out using a minimal system containing Pol I purified from yeast, recombinant CF and Rrn3 purified from E. coli in the absence (-) or presence of increasing amounts (grey triangle) of Net1 CBP , Net1ΔCTR CBP , or CTR CBP , which was either mock treated or incubated with λ-PP as described in the legend to Fig 6E. Plasmid 435 was linearized with BstZ17I and was used as Pol I promoter carrying template DNA which should yield a 591bp run-off transcript. Radiolabeled RNA was isolated and separated by 6% acrylamide-urea gel electrophoresis. An autoradiograph is shown. The position the specific transcript is indicated on the right.  Fig 7C. B) Secondary structure prediction by the Disorder Prediction Meta-Server (http://www-nmr.cabm.rutgers.edu/ bioinformatics/disorder/) for yeast Net1 (amino acids (aa) 1-1189; GI: 1023942937) and hUBF1 (aa 1-764; GI: 7657671). The disorder consensus model was created using 6 public structure prediction tools. The Disorder consensus (1 = disordered, 0 = structured) is plotted against the amino acid sequence of full-length yeast Net1 (top), and full length hUBF1. Scaled cartoons of Net1 and hUBF1 are depicted below the respective graphs with symbols described in the legends to Fig 1B and Fig 7A, respectively. A red dotted rectangle frames the regions shown in the pairwise sequence alignment in Fig 7A. C) Haploid yeast strains (y4066; y4068; y4070; y4072; y4074) expressing the indicated MN HA fusion proteins were subjected to ChEC analyses as described in Fig 4. (TIF) S1 Table. Oligonucleotides used in this study. The respective oligonucleotide names/database numbers, sequences, as well as a short description are given. (XLSX) S2  Table. Templates used for Southern blot probe generation. (XLSX) S1 Dataset. Raw data of growth analyses using automated OD measurement by TECAN reader systems. Growth analysis of liquid cultures was performed in a TECAN reader system in 96-well culture plates at 30˚C under occasional shaking. The optical density at 600/612nm of each culture was determined in 15 minute time intervals. The coordinates of the 96-well plate, strain names/database numbers, cultivation times and OD measurements are given. The OD was subsequently corrected for the blank, which was determined individually in each experiment. The OD or the logarithm of the OD were plotted against the culture time. OD values in the exponential growth phase were used to calculate doubling times presented in Figs 2B, 3B, 6D and 8D. (XLSX) S2 Dataset. Raw data of quantitative PCR analysis performed in the RotorGene Q system. Representative ChIP data of three individual experiments. CT values for each sample are given. The fluorescence measurements plotted against the cycle number and melting curve analyses are shown as diagrams. Additionally, values for "take-off" and "amplification" which are used by the comparative quantitation module of the RotorGene software to compute the "comparative concentration" used to calculate the % IP values and standard deviation errors are shown. (XLSX)