Abundance of Ribosomal RNA Gene Copies Maintains Genome Integrity

See allHide authors and affiliations

Science  05 Feb 2010:
Vol. 327, Issue 5966, pp. 693-696
DOI: 10.1126/science.1179044


The ribosomal RNA (rDNA) gene repeats are essential housekeeping genes found in all organisms. A gene amplification system maintains large cluster(s) of tandemly repeated copies in the chromosome, with each species having a specific number of copies. Yeast has many untranscribed rDNA copies (extra copies), and we found that when they are lost, the cells become sensitive to DNA damage induced by mutagens. We show that this sensitivity is dependent on rDNA transcriptional activity, which interferes with cohesion between rDNA loci of sister chromatids. The extra rDNA copies facilitate condensin association and sister-chromatid cohesion, thereby facilitating recombinational repair. These results suggest that high concentrations of heavily transcribed genes are toxic to the cells, and therefore amplified genes, such as rDNA, have evolved.

For genes that encode RNA products, such as tRNA and rRNA, an increase in expression cannot be facilitated by a translation amplification step. Thus, cells will often have multiple copies of such genes. This is especially true for the ribosomal RNA (rDNA) genes, as the product, ribosomal RNA, accounts for ~60% of total RNA in the cell. rDNA exists in high copy number in most eukaryotic cells (fig. S1). The yeast Saccharomyces cerevisiae has ~150 rDNA copies per haploid on chromosome XII (chr XII) in a tandem array (fig. S2A) (1). Repetitive genes tend to lose copies through homologous recombination among the repeats. The rDNA has a gene amplification system, thought to be present in most eukaryotes, that functions to keep high copy numbers. When copy number is reduced by recombination, DNA double-strand breaks (DSBs) are induced at the replication fork blocking (RFB) site and repaired by unequal sister-chromatid recombination, which increases rDNA copy number by rereplication of some repeats (fig. S2B) (2). Despite this highly regulated amplification system, about half the copies are not transcribed in yeast, and some plants have several thousands of rDNA copies that are likely not transcribed. We investigated the function of these extra copies in yeast.

We created four isogenic strains with different rDNA copy numbers (20, 40, 80, and 110 copies) (Fig. 1A and fig. S3). These low-copy-number strains showed normal expression levels of rRNA and grew normally in various media (fig. S4A). It was reported that in the 40-copy strain, most rDNA copies are transcribed and the polymerase density on each gene increases (3). We confirmed that the ratio of actively transcribed rDNA genes increases when the copy number reduces (fig. S4B). We found that strains with fewer rDNA copies are more sensitive to DNA damaging agents such as methyl methanesulfonate (MMS) and ultraviolet (UV) radiation (Fig. 1B). To study the basis of this sensitivity, we used the 20- and 110-copy strains to represent low and wild-type copy number conditions, respectively.

Fig. 1

Low-rDNA-copy strains are sensitive to DNA damage by MMS and UV. (A) Detection of the length of chr XII in different rDNA copy number strains by CHEF. The left panel shows the ethidium bromide (EtBr)–stained chromosome profile, and the right panel is hybridization with chr XII probe. Positions of chr XII and approximate rDNA copy numbers are indicated. (B) Spot test for DNA damage sensitivity of yeast cells with different rDNA copy numbers. YPD, yeast extract, peptone, and dextrose.

We examined whether the up-regulation of rDNA transcription affects DNA damage sensitivity in the low-copy strain. We blocked rDNA transcription by knocking out the genes for Rpa135p, which is the RNA polymerase I (pol I) subunit, and Rrn3p, which recruits RNA pol I to the 35S rRNA gene promoter. Both strains (110- and 20-copy) showed similar levels of MMS sensitivity (Fig. 2A), indicating that sensitivity of the low-rDNA-copy strain to MMS is dependent on pol I transcription.

Fig. 2

Low-copy strain shows rDNA transcription–dependent DNA damage sensitivity and rDNA replication defects. (A) DNA damage sensitivity is dependent on 35S rDNA transcription through RNA pol I (spot test). (B) 2D gel electrophoresis to detect rDNA replication “Y-arc” and recombination “spike” intermediates. (C) The ratio of recombination intermediates (X-shaped molecules) to total intermediates (X- and Y-shaped molecules) was calculated. Standard deviations are shown. (D and E) Replication status in 110- and 20-copy strains (CHEF). Cells were released from α factor into YPD containing 0.008% MMS. Chromosome pattern was detected by EtBr staining (D) and Southern hybridization with chr XII–specific probe (E). Triangles indicate positions of chr XII in the well (with asterisk) and gel. YPGal, yeast extract, peptone, and galactose.

We found no reduction in total RNA production in the 20-copy strain compared with the 110-copy strain, even after treatment with MMS (fig. S4C). Because total RNA is mostly composed of rRNA, this suggests that pol I transcription is unaffected by the damaged template in the 20-copy strain. Alternatively, heavy rDNA transcription in the 20-copy strain may prevent MMS-induced DNA lesions from being repaired, such that when replication forks encounter these DNA lesions, they stall and collapse. To investigate this possibility, we monitored DNA replication fork status in S-phase cells using two-dimensional (2D) gel electrophoresis (4, 5). Cells were synchronized in G1 phase by α factor and released into MMS-containing medium. In both the 110- and 20-copy strains, the Y-arc signal that corresponds to Y-shaped replication intermediates appears after 30 min (Fig. 2B) and then declines in both strains. In contrast, the spike signal that corresponds to X-shaped recombination intermediates (Holliday structure) increases and remains present through S phase in the 20-copy strain but does not change in the 110-copy strain (Fig. 2, B and C). The accumulation of recombination intermediates in the low-copy strain was only observed in MMS-containing medium (fig. S5A), which suggests that MMS-induced DNA damage triggers increased recombination in the low-copy strain.

We investigated the replication status of the rDNA array using contour-clamped homogeneous electric field gel electrophoresis (CHEF) (6) (Fig. 2, D and E). In the 110-copy strain, as the cells synchronously progressed through S phase, increasing quantities of chr XII bearing the rDNA remained in the well due to the presence of unresolved replication intermediates. Upon completion of replication, chr XII molecules were once again able to enter the gel. In the 20-copy strain, there was very little re-entry of chr XII into the gel, even 120 min after release, and more broken chromosome fragments were observed (Fig. 2E, lanes 60 to 120 min). Thus, completion of chr XII replication is defective in the 20-copy strain in the presence of MMS. Chromosome II also failed to re-enter the gel in the 20-copy strain, but this defect is not as severe as that seen for chr XII (fig. S6A). Analysis by fluorescence-activated cell sorting (FACS) also confirms that total DNA synthesis is not greatly affected (fig. S6B). Consistent with greater rDNA instability in the 20-copy strain, large budded cells accumulated, and the checkpoint effector kinase, Rad53p remained phosphorylated (fig. S7). Taken together, these results suggest that MMS damage inhibits rDNA replication in the low-copy strain and prevents the cells from entering G2.

To analyze the rDNA replication defect in the low-rDNA-copy strains, we looked for mutants in which copy number–dependent sensitivity is abolished. Such mutants are expected to be sensitive to MMS, therefore we examined known MMS-sensitive mutants (7). Most of the mutants still showed the difference in MMS sensitivity in the low- and high-copy-number-strains (fig. S8A). However, rad52, srs2, sgs1, mre11, rad50, and xrs2 did abolish copy number–dependent MMS sensitivity (fig. S8B). These genes are all involved in recombination repair. The accumulation of recombination intermediates in the 20-copy strain is completely abolished in a rad52 mutant (fig. S5B). These data suggest that recombination repair of the rDNA is compromised in the low-copy strain as a result of heavy pol I transcription.

The mutants that abolish rDNA copy number–dependent MMS sensitivity also show synthetic lethality with cohesion mutants (8), which suggests that their gene products collaborate with sister-chromatid cohesion in the recombination repair process. Cohesion is known to be required for sister-chromatid recombination in the repair of double-strand breaks and broken replication forks (911). Up-regulation of rDNA transcription in the low-copy-number strain may prevent cohesion, and this might compromise recombinational repair of broken forks. To test this idea, we inserted the lacO array into the rDNA and its flanking sites in the 110- and 20-copy strains (fig. S9) expressing the lacI-GFP (green fluorescent protein) fusion protein (Fig. 3A) (12). The sister chromatids of the rDNA (RDN) in the 20-copy strain are prominently separated (two dots) compared with those in the 110-copy strain at G2/M phase, indicating failure of cohesion, whereas such a difference is not observed outside the rDNA regions (CEN and TEL) (Fig. 3B). Moreover, in the low-copy strains, the ratio of separated to cohesed rDNA is negatively correlated with copy number (Fig. 3C). This difference in sister-chromatid separation by copy number disappears when 35S transcription is reduced (Fig. 3D).

Fig. 3

Low-copy strains show defects in sister-chromatid cohesion at the rDNA (cohesion assay). (A) Visualization of sister-chromatid cohesion in the rDNA by GFP Lac repressor staining at G2/M. (B) Time-course analysis of sister-chromatid cohesion in the rDNA and flanking regions (fig. S9, cohesion assay). (C) The loss rate of cohesion (RDN) is inversely correlated with rDNA copy number. (D) The cohesion defect is suppressed by reduction of pol I transcription. rpa135 temperature sensitive (rpa135-1) mutants were analyzed. At least 800 cells were scored for each strain, and standard deviations are shown [(C) and (D)]. (E) LacIfull-GFP artificially tethers the sister chromatids (RDN) in the 20-copy strain (cohesion assay). (F) DNA damage sensitivity in the low-copy strain is suppressed by artificial tethering of sister chromatids. BFI, bright field image.

To investigate whether sister-chromatid separation is a direct cause of MMS sensitivity, we artificially tethered the sister chromatids together in the 20-copy strain using the full-length lacI protein (lac-Ifull) (Fig. 3E) (12, 13), which binds to itself and the lacO arrays inserted at all of the rDNA repeating units (see Supporting Online Material). As shown in the spot test, the viability of the tethered strain in MMS-containing medium was clearly higher than that of the control nontethered strain (Fig. 3F, No IPTG). Moreover, addition of isopropyl-β-d-thiogalactopyranoside (IPTG), which inhibits the association of lac-Ifull with the lacO arrays, reduced the viability of the lac-Ifull strain to the control nontethered level (Fig. 3F, IPTG). These results indicate that sister-chromatid separation in the low-copy strain increases sensitivity to damage.

We tested cohesin association with the rDNA in the 110- and 20-copy strains and found no significant difference (fig. S10A). The condensin complex contributes to rDNA condensation and cohesion (14, 15), and its association with rDNA is diminished by 35S rDNA transcription (16, 17). In S phase, condensin association with the 35S rDNA in the 20-copy strain is clearly reduced compared with the 110-copy strain (Fig. 4, A and B). MMS treatment does not affect this association (fig. S10B). These results suggest that increased rDNA transcription in the low-copy strain inhibits condensin association during replication and that this association is not affected by DNA damage. We also investigated the contribution of condensin to cohesion in the rDNA by measuring the cohesion efficiency of a condensin mutant (ysc4-1) (18). In the high-copy strain, this mutation resulted in an increased level of separation, but there was no change in the low-copy strain (Fig. 4C), which suggests that condensin is required for cohesion of the rDNA only in the high-copy strain. Moreover, mutation of condensin in the 110-copy strain increased MMS sensitivity (Fig. 4D), consistent with previous results from S. pombe (19). In contrast, in the 20-copy strain, the condensin mutation had little effect on MMS sensitivity, which suggests that condensin is involved in damage repair in the high-rDNA-copy strain. Taken together, these results indicate that in the low-copy strain condensin association with the rDNA is reduced, resulting in a reduced ability to repair DNA damage.

Fig. 4

A condensin defect causes compromised sister-chromatid cohesion and DNA damage sensitivity. (A) Condensin association with the rDNA is reduced in the low-copy strain in S phase [as determined by chromatin immunoprecipitation (ChIP) assays]. Eight regions in the rDNA were analyzed by polymerase chain reaction (PCR) (left). Values of immunoprecipitated DNA (IP) were normalized by total DNA (Input) and “base,” the PCR control region (located between 5S and 35S rDNA) (right). (B) Test of the ChIP quantitative PCR assay. PCR samples were diluted twofold to check that the cycle was in the quantitative range (left). Quantification of condensin association with the rDNA is shown on the right. CEN5 is a non-rDNA control. (C) Condensin is required for sister-chromatid cohesion in the rDNA in the high-rDNA-copy strain (cohesion assay). (D) Condensin acts in DNA damage resistance in the high-rDNA-copy strain but not in the low-rDNA strain (spot test). The CHEF analysis confirms rDNA copy numbers. Standard deviations are shown [(A) to (C)]. IGS, intergenic spacer; YCS, yeast condensin subunit.

Our results suggest that multiple copies of rDNA are required to reduce rDNA transcription and allow efficient replication-coupled recombination repair by facilitating condensin association and sister-chromatid cohesion. This indicates that condensin self-assembly functions not only in DNA compaction but also to attach the rDNA arrays from the sister chromatids to each other. These results also provide a clue to evolution. Evolution of rDNA copy number may be related to cell size. Bacteria are considerably smaller than eukaryotic cells, and they have only a few copies of the rDNA dispersed throughout the genome and do not have an rDNA amplification system (20). Bigger cells needed more ribosomes and rDNA transcription. This increased rDNA transcription would have been toxic due to greater sensitivity to DNA damage caused by environmental factors such as ultraviolet radiation and x-rays, selecting for cells that can maintain multiple rDNA copies, and resulting in the evolution of the rDNA amplification system.

Supporting Online Material

Materials and Methods

Figs. S1 and S10

Table S1


References and Notes

  1. We thank M. Nomura (University of California, Irvine), H. Araki National Institute of Genetics, Japan (NIG), A. Murray (Harvard University), and K. Shimada (Friedrich Miescher Institute, Switzerland) for kindly providing plasmids and strains. We thank A. R. D. Ganley (Massey University) and T. Iida (NIG) for critical reading and comments on the manuscript. This work was supported in part by grants-in-aid for Scientific Research (17080010, 21247003, and 17370065 to T.K.) from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT), Japan.

Stay Connected to Science

Navigate This Article