Ixr1 is required for the expression of the ribonucleotide reductase Rnr1 and maintenance of dNTP pools

Abstract

The Saccharomyces cerevisiae Dun1 protein kinase is a downstream target of the conserved Mec1-Rad53 checkpoint pathway. Dun1 regulates dNTP pools during an unperturbed cell cycle and after DNA damage by modulating the activity of ribonucleotide reductase (RNR) by multiple mechanisms, including phosphorylation of RNR inhibitors Sml1 and Dif1. Dun1 also activates DNA-damage-inducible genes by inhibiting the Crt1 transcriptional repressor. Among the genes repressed by Crt1 are three out of four RNR genes: RNR2, RNR3, and RNR4. The fourth RNR gene, RNR1, is also DNA damage-inducible, but is not controlled by Crt1. It has been shown that the deletion of DUN1 is synthetic lethal with the deletion of IXR1, encoding an HMG-box-containing DNA binding protein, but the reason for this lethality is not known. Here we demonstrate that the dun1 ixr1 synthetic lethality is caused by an inadequate RNR activity. The deletion of IXR1 results in decreased dNTP levels due to a reduced RNR1 expression. The ixr1 single mutants compensate for the reduced Rnr1 levels by the Mec1-Rad53-Dun1-Crt1-dependent elevation of Rnr3 and Rnr4 levels and downregulation of Sml1 levels, explaining why DUN1 is indispensible in ixr1 mutants. The dun1 ixr1 synthetic lethality is rescued by an artificial elevation of the dNTP pools. We show that Ixr1 is phosphorylated at several residues and that Ser366, a residue important for the interaction of HMG boxes with DNA, is required for Ixr1 phosphorylation. Ixr1 interacts with DNA at multiple loci, including the RNR1 promoter. Ixr1 levels are decreased in Rad53-deficient cells, which are known to have excessive histone levels. A reduction of the histone gene dosage in the rad53 mutant restores Ixr1 levels. Our results demonstrate that Ixr1, but not Dun1, is required for the proper RNR1 expression both during an unperturbed cell cycle and after DNA damage.

Figure 1. Mec1-Rad53-Dun1–dependent regulation of S. cerevisiae ribonucleotide reductase.

The activated Dun1 kinase relieves inhibition of RNR by targeting the transcriptional repressor Crt1(Rfx1) and protein inhibitors Sml1 and Dif1.

Figure 2. dun1 ixr1 synthetic lethality is rescued by increased RNR activity.

(A) Schematic representation of the Ixr1 protein. HMG boxes are shown in blue and polyQ regions in black. (B) Alignment of homologous HMG proteins close to the conserved S366. (C) Tetrad analysis demonstrating that deletion of SML1 rescues the dun1Δ ixr1-S366F synthetic lethality. The dun1Δ ixr1-S366F strain (TOY544) carries a DUN1-containing plasmid, pK503, which confers a red color; this plasmid is lost in ixr1-S366F dun1Δ sml1Δ colonies (TOY544×TOY588). (D) Tetrad analysis demonstrating that deletion of SML1 rescues the ixr1Δ dun1Δ synthetic lethality (TOY604×TOY527). (E) Tetrad analysis demonstrating that a low-copy RNR1 vector rescues the ixr1Δ dun1Δ synthetic lethality. The diploid strain ixr1/IXR1 dun1/DUN1 (TOY527×TOY603) was transformed with pRS316 and pBJ6 (pRS316-RNR1), sporulated and tetrad analysis was performed. (F) Overexpression of RNR1 rescues the ixr1-S366F dun1Δ synthetic lethality. The pK503 plasmid, containing the DUN1 gene and conferring red color, is lost on YP-Gal medium in the ixr1-S366F dun1Δ strain transformed with pESC-URA-pGAL1-RNR1 plasmid (left panel, sectoring phenotype), but not when transformed with pESC-URA-pGAL1 (right panel). (G) Low-copy RNR1 vector rescues the ixr1-S366F dun1Δ synthetic lethality. The pK503 plasmid, containing the DUN1 gene and conferring red color, is lost in the ixr1-S366F dun1Δ strain transformed with the pBJ6 plasmid (left panel, sectoring phenotype), but not when transformed with pRS316 (right panel). (H) Overexpression of RNR3 rescues the ixr1-S366F dun1Δ synthetic lethality. The pK503 plasmid, containing the DUN1 gene and conferring red color, is lost in the ixr1-S366F dun1Δ strain transformed with pBAD79 plasmid (left panel, sectoring phenotype), but not when transformed with pRS414 (right panel).

Figure 3. Deletion of IXR1 leads to increased Rnr3 and Rnr4 levels and decreased Sml1 levels.

(A) Western blot analysis of Rnr3-HA and Rnr4 levels in wild-type (AC447-2A), ixr1-S366F (TOY619), and ixr1Δ (TOY621) strains before and after 2 hours treatment with 0.2 mg/L 4-nitroquinoline 1-oxide (4-NQO), 200 mM HU, or 0.02% methyl methanesulfonate (MMS). Rnr3 and Rnr4 levels were quantified in relative units (RU, levels of Rnr3 or Rnr4 divided by the levels of tubulin in corresponding sample) as described in Materials and Methods. ND – not detected. (B) Western blot analysis of Rnr2 levels in wild-type (AC447-2A) and ixr1Δ (TOY621) strains before treatment and after 2 hours treatment with 0.2 mg/L 4-NQO, 0.02% MMS, or 200 mM HU. (C) Western blot analysis of Sml1 levels in wild-type (W1588-4C) and ixr1Δ (TOY736) strains before treatment and after 2 hours treatment with 0.02% MMS. (D) Western blot analysis of Rad53 phosphorylation status in wild-type (W1588-4C) and ixr1Δ (TOY736) strains before treatment and after 2 hours treatment with 0.02% MMS. (E) Deletion of RAD53 but not of SML1 or DUN1 abolishes the upregulation of Rnr3 and Rnr4 levels in ixr1Δ. Western blot analysis of Rnr3-HA and Rnr4 levels. The following strains were analyzed: wt (AC447-2A), ixr1Δ (TOY732), ixr1Δ sml1Δ (TOY778), ixr1Δ dun1Δ sml1Δ (TOY772), and ixr1Δ rad53Δ sml1Δ (TOY781).

Figure 4. Deletion of IXR1 leads to decreased dNTP levels.

(A) dNTP levels are decreased in ixr1Δ (TOY736), increased in ixr1Δ sml1Δ (TOY714) compared to wild-type (W1588-4C), but lower than in sml1Δ (U952-3B). Values shown are the average from two independent experiments with the minimum and maximum values represented as error bars. (B) dNTP levels increase in ixr1Δ (TOY736, black triangles) during the treatment with 0.2 mg/L 4-NQO, but less than in a wild-type strain (W1588-4C, open circles). Values shown are the average from two independent experiments with the minimum and maximum values represented as error bars. (C) ixr1Δ (TOY736) has a higher frequency of petite formation than a wild-type strain (W1588-4C). (D) ixr1Δ (TOY736) is more sensitive to HU than a wild-type strain (W1588-4C).

Figure 5. Rnr1 levels are reduced in ixr1 after DNA damage.

(A) Western blot analysis of Rnr1 levels in wild-type (wt) (W1588-4C), ixr1-S366F (TOY734), and ixr1Δ (TOY736) strains before and after 2 hours treatment with 0.2 mg/L 4-NQO or 200 mM HU. (B) Quantification of Rnr1 levels in wild-type (wt) (W1588-4C), ixr1Δ (TOY736) and ixr1-S366F (TOY734) strains before and after 2 hours treatment with 0.2 mg/L 4-NQO or 200 mM HU. Tubulin was used as the internal control (see Materials and methods). Error bars represent standard error of the mean (SEM). (C) Lower panel, dynamics of Rnr1 decrease and Rnr4 increase in wt (W1588-4C, lanes marked with “w”) and ixr1Δ (TOY736, lanes marked with “i”) strains during a 12-hour time-course incubation with 0.2 mg/L 4-NQO. Upper panel, the corresponding flow-cytometric histograms. (D) Proliferation curves for the experiment in Figure 4C.

Figure 6. Ixr1 regulates RNR1 promoter activity.

(A) β-Galactosidase assay of RNR1, RNR3, and RNR4 promoter activity in wild-type (W1588-4C) and ixr1Δ (TOY736) strains. Promoters of the analyzed genes were fused with the lacZ repoter gene and the respective plasmids were transformed in the wild type and ixr1Δ strains. β-Gal units were quantified as described in Materials and Methods. (B) Analysis of DNA associated with Ixr1 was performed by chromatin immunoprecipitation (ChIP) followed by qPCR using locus-specific primers for the RNR1 promoter (pRNR1), the DSF2 promoter (pDSF2) and the ACT1 gene. ChIP was performed with IXR1 (W1588-4C) and IXR1-9MYC (TOY836) strains using anti-Myc antiserum 9E10 or mock-antiserum. For pRNR1, ChIP was performed both without and with addition of 4-NQO (0.2 mg/L). Y axis represents the amount of precipitated DNA relative to the input DNA. Values shown are the average from two independent experiments with the minimum and maximum values represented.

Figure 7. Elevation of Rnr1 levels in response to DNA damage requires MEC1, RAD53, and IXR1, but not DUN1.

(A) Western blot analysis of Rnr1 and Rnr4 levels. The following strains were incubated with 0.2 mg/L 4-NQO for 2 hours: wt (W1588-4C), ixr1Δ (TOY736), dun1Δ sml1Δ (TOY728), rad53Δ sml1Δ (TOY782), mec1Δ sml1Δ (TOY711), dun1Δ rad53Δ sml1Δ (TOY786), and dun1Δ mec1Δ sml1Δ (TOY774). Rnr1 and Rnr4 levels were quantified in relative units (RU, levels of Rnr1 or Rnr4 divided by the levels of tubulin in corresponding sample) as described in Materials and Methods. (B) Western blot analysis of λ-phosphatase treated extracts from Ixr1-Ha (TOY655) and Ixr1-S366F-Ha (TOY650) strains. (C) Western blot analysis of Ixr1 levels in wt (W1588-4C), dun1Δ sml1Δ (TOY728), rad53Δ sml1Δ (TOY782), mec1Δ sml1Δ (TOY711), rad53Δ dun1Δ sml1Δ (TOY786) and mec1Δ dun1Δ sml1Δ (TOY774). (D) Western blot analysis of Ixr1 levels in wild-type (W1588-4C), hht2-hhf2Δ (TOY806), hht2-hhf2Δ sml1Δ (TOY821), rad53Δ sml1Δ (TOY782) and hht2-hhf2Δ rad53Δ sml1Δ (TOY819). Ixr1 levels were quantified in relative units (RU, levels of Ixr1 divided by the levels of tubulin in corresponding sample) as described in Materials and Methods. (E) Mec1-Rad53-Dun1-dependent regulation of S. cerevisiae ribonucleotide reductase. Expression of RNR1 depends on Mec1, Rad53 and Ixr1, but does not depend on Dun1 or Crt1.