Ixr1 Regulates Ribosomal Gene Transcription and Yeast Response to Cisplatin

Ixr1 is a Saccharomyces cerevisiae HMGB protein that regulates the hypoxic regulon and also controls the expression of other genes involved in the oxidative stress response or re-adaptation of catabolic and anabolic fluxes when oxygen is limiting. Ixr1 also binds with high affinity to cisplatin-DNA adducts and modulates DNA repair. The influence of Ixr1 on transcription in the absence or presence of cisplatin has been analyzed in this work. Ixr1 regulates other transcriptional factors that respond to nutrient availability or extracellular and intracellular stress stimuli, some controlled by the TOR pathway and PKA signaling. Ixr1 controls transcription of ribosomal RNAs and genes encoding ribosomal proteins or involved in ribosome assembly. qPCR, ChIP, and 18S and 25S rRNAs measurement have confirmed this function. Ixr1 binds directly to several promoters of genes related to rRNA transcription and ribosome biogenesis. Cisplatin treatment mimics the effect of IXR1 deletion on rRNA and ribosomal gene transcription, and prevents Ixr1 binding to specific promoters related to these processes.

The high-mobility group (HMG) proteins, present in almost all metazoans and plants, were discovered as nuclear factors over 40 years ago 1 . The HMG-box that characterizes the HMGB subfamily 2 comprises 3 α-helices folded into an L-shaped configuration, in which the concave surface binds to the minor groove of DNA 3 . They act in the nucleus as non-histone architectural-chromatin proteins, having regulatory functions in replication, transcription and DNA repair 4 . HMGB proteins also bind with high affinity to distorted DNA structures, such as 4-way junctions, bulges, kinks, and modified DNA containing cisplatin adducts 5 . Ixr1 (Intrastrand cross (X)-link recognition, formerly called Ord1, for Oxygen/oxidase regulation defective) is a Saccharomyces cerevisiae HMGB protein that regulates the transcription of genes involved in the response to normoxia-hypoxia changes [6][7][8][9][10] . Ixr1 also binds to cisplatin-DNA adducts with high affinity 11 . A model to explain the dual role of Ixr1 in transcriptional regulation and recognition of cisplatin-DNA adducts has recently been proposed 12 .
Cisplatin is used in cancer chemotherapy 13 , however cancerous cells usually acquire resistance against the drug soon after the initial treatment, thus limiting its effectiveness 14 . The molecular mechanisms of cisplatin cytotoxicity and acquired resistance in mammals have been thoroughly reviewed 15 . Yeast has been used as an easy-to-handle eukaryotic model to find genes related to cisplatin responsiveness and IXR1 deletion leads to a hyperresistance [16][17][18] . IXR1 mutation also favors the rate of spontaneous mutagenesis mediated by replication errors 19 . The hypothesis that Ixr1 and other HMG-domain proteins might block repair of the major cisplatin-DNA adducts in vivo, thus inducing cell death, was postulated over 20 years ago 20 . It is supported by the evidence that IXR1 deletion does not increase resistance of S. cerevisiae cells to cisplatin that already have mutations in the RAD genes related to DNA-repair 20 . A more recent model suggests that elimination of Ixr1 pre-activates the genome integrity checkpoint above basal levels, thereby increasing DNA repair and cisplatin resistance 21 . Ixr1 is also required for the maintenance of an adequate supply and balance of dNTPs for DNA synthesis and repair 22 .
We have analyzed transcriptomes to compare the regulatory roles of Ixr1 in absence or presence of cisplatin. Ixr1 regulates in yeast other transcriptional factors, which respond to external stimuli and control cell growth and proliferation. A number of approaches -qPCR, ChIP to CHIP, and quantification of 18S and 25S rRNAs -confirm the function of Ixr1 in the control of ribosome biogenesis. Connections between this control, the TOR signaling pathway and the effects of cisplatin are discussed.  effectors, or by other post transcriptional mechanisms controlling their mRNA levels. Alternatively, these genes could be considered as "transient targets" of Ixr1, defined as those regulated, but without detectable binding 41 .
qPCR measurements were made to validate the transcriptional regulation exerted by Ixr1 on the expression of these transcription factors representing the major nodes in the regulation network shown in Fig. 3A. These data, which included the analysis of genes SFP1, ABF1, TEC1, SOK2, UME6, and DAL81, confirmed the findings (Fig. 3B). Also included in this analysis was the CRF1 gene that encodes a co-repressor protein of ribosomal protein (RP) genes and is upregulated in the ixr1Δ strain (Table S1) and therefore under the negative control exerted by Ixr1. Upregulation in the ixr1Δ strain was also confirmed by qPCR (Fig. 3B).
Ixr1 control on transcription of Sfp1, rRNA, and genes involved in ribosome biogenesis, functionally affects rRNA levels. Synthesis of ribosomal components and their assembly is closely associated with cell growth and proliferation 27 ; 90% of total cellular transcription is used during ribosome biogenesis of rapidly growing cells 56 . In S. cerevisiae, transcription is the major level of regulation of ribosome biogenesis 57,58 , involving all 3 nuclear RNA polymerases. As described above, SFP1 is downregulated in the W303-ixr1Δ strain (Table S3). Sfp1 is a zinc finger protein that regulates transcription of ribosomal proteins and other genes related to ribosome biogenesis 45 , as also responses to DNA-damage, nutrient availability, cell cycle progression 46 and cisplatin resistance 59 . It binds DNA directly at highly active RP genes and probably indirectly through Rap1 protein at others 60 .   To test whether the transcriptional regulation of genes related to rRNA transcription had a functional significance, we first measured 18S and 25S rRNAs levels by fluorescence assay (Fig. 2C), which showed a significant decrease in the ixr1∆ strain in reference to W303. The levels of 18S, 5.8S, and 25S rRNAs were also measured, as well as those of their pre-processed forms 35S, 27S, 20S, by using qPCR and specific primers for each form (Table S7). The levels of these transcripts were always significantly higher in the wild type strain W303 than in the ixr1∆ strain (Fig. 2C).
Effect of IXR1 deletion on transcriptional response to cisplatin treatment. The effects of cisplatin treatment upon the S. cerevisiae transcriptome have been previously reported 61 . The major functional group over-represented among downregulated genes is related to ribosome biogenesis, including genes involved in the maturation of SSU-rRNA from tricistronic rRNA transcript, or those participating in ribosomal small subunit assembly and/or rRNA export from the nucleus 61 . We have directly measured the cisplatin effect on the levels of 18S, 5.8S, and 25S rRNAs, and their pre-processed forms (35S, 27S, 20S) in the wild type strain W303, and results obtained ( Figure S2) also confirmed the decrease observed in the previous transcriptome assay 61 .
To more fully understand the role of Ixr1 in the response to cisplatin, we used transcriptome analysis to see how the deletion of IXR1 affects this response. Significant changes as a consequence of cisplatin treatment occurred in both W303 and ixr1∆ strains (Fig. 4A). The transcriptional activation of genes related to metabolism of sulfur compounds, including cysteine and methionine, is stronger in the ixr1Δ strain than in the W303 strain (Fig. 4B). However, expression of most genes related to ribosome biogenesis remain unchanged or with low change in the ixr1∆ strain treated with cisplatin ( Fig. 4C and D), indicating that Ixr1 is necessary in the ribosomal-gene response to this compound.
DNA binding of Ixr1 in the presence of 600 μM cisplatin was analyzed with the same criteria used for the analysis of the data where the compound had not been used. Delimiting upstream regions up to −1000 bp from the ATG, 237 peaks were obtained. Only 85 Ixr1-binding peaks were on the same promoters found in the absence of cisplatin (Fig. 5A). A pool of 152 promoter peaks was exclusively detected in the presence of cisplatin treatment. Among these genes, the GO term "substrate-specific transmembrane transporter activity" [GO:0022891] is significantly enriched (p < 0.05) and includes 20 matches. This group includes hexose and amino acid transporters and other genes that have been related to detoxification processes. The latter group includes FUI1, involved in the transport of the cytotoxic nucleoside analog 5-fluorouridine 62 or PUT4, related to the transport of the toxic proline analog, azetidine-2-carboxylate 63 ; FTR1 and ZRT1, related to the transport of metals 64,65 ; TPO1, involved in exporting spermine and spermidine from the cell during oxidative stress 66 ; finally and remarkably, QDR2 encoding a protein that exports copper and also a range of other compounds, and which is required for resistance to cisplatin and other drugs 67 .
As previously explained, the effect of cisplatin on the transcription of genes related to ribosome biogenesis diminishes in the absence of Ixr1 ( Fig. 4C and D). Peaks that are not formed in presence of cisplatin include important regulators for rRNA or ribosome biogenesis, such as TOD6, ENP1, RRP7, NMD3, HCA4, UBP10, ECM23 or RSA4 27 . The consensus for Ixr1 binding is found among the binding peaks seen in absence of cisplatin, whereas it is not found in the analysis of peaks exclusively formed after cisplatin treatment (Fig. 5C). Remarkably, up to 49% of the Ixr1 promoter binding peaks seen in presence of cisplatin contain a consensus sequence related to specific binding of the Rsc30 protein (Fig. 5C).

Discussion
Our work demonstrates that Ixr1 controls ribosome biogenesis by direct binding to promoters of specific genes that regulate rRNA, RP and RiBi gene expression, but also by indirect regulation of specific transcriptional regulators, such as Sfp1 or Crf1; in these cases without observing direct binding of Ixr1 to their promoters. As a result of both direct and indirect mechanisms, IXR1 deletion reduces 18S and 22S rRNA levels.
In eukaryotes, the target of rapamycin (TOR) signaling pathway promotes anabolic functions necessary for cell growth, while suppressing other catabolic processes, such as autophagy 68 . There are two effectors, the TOR complexes 1 and 2 (TORC1 and TORC2), which have functional specialization. TORC1 is sensitive to rapamycin, is related to nutrient signaling, and controls cell proliferation and ribosome biogenesis; TORC2, which is rapamycin insensitive, is associated with the control of actin cytoskeleton and cell cycle progression 69 . It is also involved in cell wall integrity and sphingolipid metabolism 70 . Considering that TOR activation promotes ribosomal biogenesis and Ixr1 controls transcription of genes related to this process, the intriguing question is how Ixr1 transcriptional control overlaps with the TOR signaling pathway. Indeed, other yeast HMG proteins are involved in TOR signaling, albeit by epigenetic mechanisms 71 .
Comparison of Ixr1 targets and TOR signaling pathway components shows overlap ( Figure S1). It has also been suggested that TORC1 is activated by an abundance of leucine 72 ; remarkably, Ixr1 is necessary for active transcription of genes encoding enzymes involved in the synthesis of leucine, isoleucine and valine 10 , which might contribute to the deactivation of TORC1 signaling in the absence of Ixr1. The TOR pathway controls DNA damage responses by regulating dNTP production 45,73,74 , which again connects to Ixr1 function, since Ixr1 regulates dNTP pools 22 . Although nitrogen activates the TOR signaling pathway 75 , optimal growth also requires a carbon source. In yeast, the cAMP/PKA pathway, which works on the basis of nutrients availability, growth, proliferation, metabolism, stress resistance, aging and morphogenesis, is activated by glucose 76 . Since TOR and cAMP/PKA are connected 77,78 , changes in glucose availability may also affect the final TOR targets. Deletion of IXR1 (Fig. 4A) also affects the expression of genes encoding hexose transporters and Sok2, which is involved in the response to glucose and it is phosphorylated by PKA 50 .
From our data and in the context of published papers, a hypothetical model of cooperation between Ixr1-dependent and TOR-dependent mechanisms for maintaining ribosome biogenesis is proposed (Fig. 6). We found that when Ixr1 is functional, SFP1 is actively transcribed (Fig. 6A). The subcellular localization of Sfp1, the master regulator of Ribi and RP genes, is regulated by both the cAMP/PKA and TOR network in response to nutritional and stress inputs 45 . In growing cells not limited by nutrient availability, the TOR1 pathway is also active and therefore Sfp1, the key regulator of RiBi and RP genes, localizes in the nucleus where it activates the transcription of these regulons 44,45,79 , thus allowing ribosome biogenesis, growth and proliferation (Fig. 6A). The other point of transcriptional activation mediated by Ixr1 upon the RP genes is through repression of the co-repressor Crf1 (Fig. 6A). Fhl1, a forkhead-like protein has a dual role as activator (in association with the Ifh1 co-activator) or repressor (in association with the Crf1 co-repressor) in the transcription of the RP genes [80][81][82] . Since Fhl1 is constitutively bound to the RP gene promoters, its activity depends on the presence of Ifh1 or Crf1. In growing cells, TOR keeps Crf1 inactive in the cytoplasm by repressing Yak1 kinase, possibly via a PKA-dependent route 81 , allowing expression of RP genes (Fig. 6A). In the absence of Ixr1, expression of the SFP1 gene is reduced, as already noted (Fig. 6B). In response to carbon and nitrogen starvation, oxidative stress and  inactivation of TOR signaling, Sfp1 rapidly translocates to the cytoplasm (Fig. 6B), and loses its function 44,45,79 . After TOR inactivation, phosphorylated Crf1 in the nucleus displaces Ifh1 (Fig. 6B), thereby inhibiting transcription of RP genes 81 . Moreover, the nuclear localization of both Fhl1 and Ifh1 is influenced by Sfp1 80 . Therefore, Ixr1 might also affect indirectly the subcellular localization of Ifh1 and Crf1 by affecting Sfp1 expression in addition to its effect on CRF1 mRNA levels reported in our analysis (Fig. 6A).
After cisplatin treatment, the transcriptional activation of genes related to metabolism of sulfur compounds, including cysteine and methionine, is stronger in the ixr1Δ strain than in the W303 strain (Fig. 4B). Stimulation of a pathway of amino acid biosynthesis seems paradoxical in a situation in which the biosynthesis of ribosomal proteins is downregulated. However, specific sulfur-containing amino acids may be necessary to cope with oxidative stress, and their upregulation could favor cisplatin resistance in the null mutant by increasing chelating groups to immobilize the Pt compound, perhaps promote glutathione biosynthesis to favor anti-oxidant reactions, or favor cisplatin extrusion from the cell, through the formation of cisplatin-glutathione complexes. In support of this view, several mechanisms affecting cisplatin toxicity have been related to glutathione levels in human cells 83 .
We have shown that cisplatin treatment prevents Ixr1 binding to several promoters controlling ribosome biogenesis. In humans, the cytotoxic effect of cisplatin is attributed to diverse mechanisms, among which is reduced ribosome biogenesis 84,85 . Since cisplatin treatment mimics IXR1 deletion in the control of transcription of genes involved in ribosome biogenesis, one possible explanation is that the Ixr1 protein binds to cisplatin-DNA adducts at other locations, being displaced from these promoters. Alternatively the formation of cisplatin-DNA adducts might induce chromatin remodelling and affect specific Ixr1-DNA binding. We have observed that the 37% of the Ixr1 promoter binding peaks seen in presence of cisplatin contain a consensus sequence related to the specific binding region of the Rsc30 protein (Fig. 5C). Rsc30 forms part of the S. cerevisiae ATP-dependent remodeling complex RSC (Remodels the Structure of Chromatin) that has 17 subunits 86 . RSC modulates access to chromatin, and therefore controls DNA metabolism, including replication, transcription, recombination, and DNA repair. RSC performs a large number of different remodeling activities, including exchange or incorporation of core histones or histone variants, eviction of histones or nucleosomes and repositioning or sliding of nucleosomes 87 . Interestingly, RSC complexes are also involved in the transcription of genes related to ribosome biogenesis 88 ; also, strains lacking the subunits Rsc1 or Rsc2 are hypersensitive to a variety of DNA damaging agents [89][90][91][92] . The RSC complex participates in the inactivation of the TORC1 pathway in response to nitrogen starvation 93 . Considering all data, the existence of a putative interplay between Ixr1 binding and RSC-remodeling is a new question to be considered in future studies.
In summary, IXR1 deletion diminishes transcription of ribosomal RNAs and genes encoding ribosomal proteins, or which are necessary for ribosome assembly, by direct and indirect mechanisms. Cisplatin treatment mimics the effect of IXR1 deletion on rRNA and ribosomal gene transcription, and prevents Ixr1 binding to specific promoters of genes involved in these processes.

Methods
Cell culture and treatments. S. cerevisiae strain W303 (MATa ade2-1 can1-100 leu2-3,112 trp1-100 ura3-52) and its derivative, W303-ixr1Δ (previously described by 94 , have been used in the transcriptome experiments. S. cerevisiae strain Z1580 (MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 GAL + psi + IXR1: myc9::TRP1) was obtained from Young's lab 95 . Three biological replicas of cultures and treatments were made. Yeast cells were cultured overnight in 10 mL of complete medium (SD) containing per liter: 6.7 g of bacto-yeast nitrogen base without amino acids (Difco, Franklin Lakes, New Jersey, USA); 40 mg of the following additives (w/v): histidine, leucine, adenine, uracil, lysine and tyrosine; 10 mg arginine, methionine and threonine; 30 mg tryptophan; 60 mg phenylalanine and isoleucine; and 2% glucose. For transcriptome analyses on the following day, the media were inoculated at an initial OD 600 of 0.4 in 70 mL SD. Cells were grown in 250 mL Erlenmeyer flasks at 30 °C with agitation at 250 rpm. When cells had reached an OD 600 of 0.6, the cultures from each strain were continued in 2 × 25 mL aliquots, i.e. control and cisplatin treated. A stock solution of cisplatin at 6 mM in dimethyl sulfoxide (DMSO) was prepared and the drug added to the cultures at a final concentration of 600 μM, an equivalent volume of DMSO being added to the control cultures. The cells were kept at 30 °C with agitation at 250 rpm for 4 h in darkness to prevent the inactivation of cisplatin. The concentration of cisplatin and the time course of the treatment had previously been established in pilot experiments with the selected yeast strains.
For ChIP-on-chip experiments, the cells were inoculated at an initial OD 600 of 0.1 in 200 mL SD and grown in 1 L Erlenmeyer flasks at 30 °C with 250 rpm agitation. When cells had reached an OD 600 of 0.6, cisplatin was added to the cultures at a final concentration of 600 μM. The time-course of the treatment was applied as explained for transcriptome analyses.
RNA preparation and transcriptomic microarray analysis. RNA was extracted from a number of cells corresponding to an OD 600 of 3 with the AurumTM Total RNA Mini Kit (Bio-Rad). Concentration and purity of RNA was measured using the ratio R = A 260 /A 280 (always in the range of 1.7 < R < 2.1). RNA integrity was also measured by the RIN parameter (RNA Integrity Number) with a 2100 Bioanalyzer (Agilent Technologies, Inc. Santa Clara, CA 95051-7201USA). RIN was close to 9 in all the samples, which is considered high-quality extraction 96 .
Twelve GeneChip ® Yeast-Genome-2.0 arrays from Affymetrix Inc. (Wycombe. United Kingdom) were used and processed in the GeneChip ® System with Autoloader from Affymetrix Inc. (Wycombe, UK). We started from 10 ng total RNA from each sample for successive cDNA, aRNA generation, labeling with biotin and fragmentation using the GeneChip ® 3′ IVT Express Kit. RNA fragmentation was monitored with a 2100 Bioanalyzer (Agilent Technologies, Inc. Santa Clara, CA 95051-7201, USA), by selecting conditions producing fragments Scientific RepoRts | (2018) 8:3090 | DOI:10.1038/s41598-018-21439-1 from 35 to 200 nt, with a majority between 100-120 nt. Hybridization, washes and staining were done with the GeneChip ® HT Hybridization, Wash and Stain Kit. (Ambion, Inc. Affymetrix). The kit includes RNA Poly-A controls (lys, phe, thr and dap) from Bacillus subtilis to monitor the target labeling process, which serve as sensitivity indicators of target preparation and labeling efficiency. They also include the hybridization controls, comprised of a mixture of biotinylated and fragmented RNA of bioB, bioC, bioD (genes from the biosynthesis of biotin in Escherichia coli) and Cre (recombinase from bacteriophage P1). These controls monitor hybridization, and the washing and staining steps. Control Oligo B2 was included to provide alignment signals for image analysis.
Image capture and preliminary data analysis were carried out with Affymetrix ® Expression Console ™ software (v1.1). After RMA normalization of raw data from 3 biological repeats using the Affymetrix algorithm, the normalized data were analyzed using the web-suite Babelomics (v4.3) 97 . Statistical analyses to identify differentially expressed genes (DEGs) used the LIMMA (linear models for microarray data) test 98 . The FDR (False Discovery Rate) was used to correct values for multiple comparisons 99 . Statistical significant DEGs were considered those with FDR < 0.01 and a fold change of ≥1.4 in the comparisons. The original and normalized data from this study are uploaded in Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/info/ linking. html), the accession number of the series being GSE84569. The processed data and DEG identifications are shown in Tables S1-S4.
GeneChip ® S. cerevisiae Tiling 1.0 R arrays from Affymetrix Inc. (Wycombe,UK) were used with 3 biological repeats and processed in the GeneChip ® System with Autoloader also from Affymetrix. Control Oligo B2 was included to provide alignment signals for image analysis. Image caption and preliminary data analysis used Affymetrix ® Expression Console ™ software (v1.1). ChIP-on-chip raw data from Affymetrix GCOS software were analyzed using Affymetrix Tiling Analysis Software (TAS) v1.1.03 (http://www.affymetrix.com/support/developer/downloads/TilingArrayTools/index.affx).
A 2-sample analysis was conducted for both untreated and cisplatin treated cultures using specific immunoprecipitated DNA samples (from Ixr1-c-Myc tagged samples) as the 'experimental' group, and 2 genome-DNA fragmented and amplified samples as the 'control' group. Data were normalized using built-in quartile normalization and probe-level analysis with perfect match (PM) probes. Ixr1 protein occupancy profiles were visualized with Affymetrix Integrated Genome Browser (IGB). Interval analyses used TAS software with a minimum run of 150 bp a maximum gap of 250 bp, and a p-value cutoff of 10 −4 . Bed file conversions used UCSC (University of California Santa Cruz) tools (https://genome.ucsc.edu). *Bed file analyses used ChIpSeek tools (http://chipseek. cgu.edu.tw), as described by Chen et al. (2014).
Our original and normalized data have been uploaded into the Gene Expression Omnibus database (http:// www.ncbi.nlm.nih.gov/geo/info/linking.html); the accession number of the series is GSE101080, and processed data are available in Tables S5-S6. Data mining. Gene descriptions and comparative analyses of lists from DEGs were obtained from Yeast Mine (http://yeastmine.yeastgenome.org/yeastmine).
Scientific RepoRts | (2018) 8:3090 | DOI:10.1038/s41598-018-21439-1 Analysis of the expression by qPCR. Methods and procedures have been previously described 61 . The ECO Real-Time PCR System was used (Illumina) and calculations were made by the 2 −∆∆Ct method 106 . Three independent RNA extractions were assayed for each strain or condition. The list of primers is given in Table S7. RNA levels of the selected genes were corrected by the geometric mean of the mRNA level of TAF10, a gene previously verified to be constitutive in the assayed conditions and not affected by IXR1 deletion. HHO1 was used as the negative control, and was also unaffected in the ixr1∆ strain. A t-test was used to find statistically significant differences between control and cisplatin treated samples (p < 0.05).
Comparison of rRNA levels in W303 and ixr1Δ strains. Total RNA was extracted as previously described 61 from yeast cultures collected at logarithmic growth phase (OD 600 of ≈ 0.9-1). Two biological and 2 technical repeats of each were measured for the W303 and ixr1Δ strains. Prior to RNA analysis, RNA-containing pellets were resuspended in RNase-free water (Sigma-Aldrich) and incubated at 75 °C for 5 min to resolve secondary structures.
The relative amounts of 18S and 25S rRNAs per unit of total RNA were estimated by analyzing total RNA with an Agilent 2100 Bioanalyze with its RNA 6000 Nano kit (Agilent Technologies, Palo Alto, USA). Total RNA and RIN ratios were quantified under the Agilent eukaryotic total RNA program as previously described 96 . 18S and 25S rRNAs quantitation was calculated from the area under the peaks in reference to total RNA in the sample. Relative amounts of 5, 8S 18S, 25S rRNAs, as well as their precursor forms (35S, 27S and 20S), were measured by qRT-PCR following the procedures described in section 2.5. Specific primers were designed for each type (Table S7), and their relative positions indicated in Fig. 2C.