Histone modifications influence chromatin structure and thus regulate the accessibility of DNA to replication, recombination, repair, and transcription. We show here that the histone deubiquitinating enzyme Ubp10 contributes to the formation/maintenance of silenced chromatin at the rDNA by affecting Sir2p association.
IN eukaryotes, genomic DNA is packaged into chromatin, a nucleoprotein complex whose basic repeating unit is the nucleosome. The nucleosome is made up of 146 bp of DNA wrapped around a histone octamer consisting of two copies each of H2A, H2B, H3, and H4 (Luger 2003). Histones are subject to several post-translational modifications such as acetylation, methylation, phosphorylation, and ubiquitination (Berger 2002; Fischle et al. 2003; Shilatifard 2006). The addition/removal of chemical moieties is a dynamic process that can influence chromatin function by different mechanisms generating sites for interaction with additional proteins or affecting chromatin condensation. Contrary to acetylation, which is globally associated with active chromatin, histone ubiquitination regulates gene transcription in a positive and a negative way, depending on its genomic location (Zhang 2003; Kao et al. 2004; Osley 2004). In Saccharomyces cerevisiae many of the negative effects are observed at telomeres, HM and rDNA loci. In fact, histone H2B is monoubiquitinated at Lys123 by the ubiquitin conjugase Rad6/ligase Bre1 and this modification affects methylation of H3-Lys4 and H3-Lys79 catalyzed by the Set1 and Dot1 methyltransferases, respectively (Ng et al. 2002, 2003; Sun and Allis 2002; Osley 2004). These modifications, which are preferentially localized in euchromatin regions, prevent association of Sir proteins and restrict these factors to heterochromatin regions where they mediate silencing (van Leeuwen and Gottschling 2002; van Leeuwen et al. 2002; Santos-Rosa et al. 2004). Moreover, like histone acetylation, ubiquitination is dynamic. So far, two deubiquitinating enzymes, Ubp8 and Ubp10, have been shown to target monoubiquitinated histone H2B. They display overlapping and distinct functions (Emre and Berger 2006); in particular, the latter is involved in silencing (Kahana and Gottschling 1999; Orlandi et al. 2004). In this context, Ubp10p has been shown to localize at the silenced telomere-proximal loci, where it is responsible for a low level of H2B Lys123 monoubiquitination. Through a trans-histone regulation, Ubp10p cooperates in maintaining a low level of H3 Lys4 and Lys79 methylation, required for proper association of Sir proteins to telomeres (Emre et al. 2005; Gardner et al. 2005). Silent chromatin at chromosome ends contributes to genomic stability since in an open state the telomeres resemble double-strand breaks and elicit DNA repair/recombination activities. Silent chromatin is also a feature of the yeast rDNA locus: a locus highly susceptible to recombination due to its repetitive arrangement and unidirectional mode of DNA replication. Both positive and negative regulatory factors assure an accurate control in the recombination levels of rDNA (Defossez et al. 1999; Kaeberlein et al. 1999; Ivessa et al. 2000; Johzuka and Horiuchi 2002; Versini et al. 2003; Weitao et al. 2003; Blander and Guarente 2004). Sir2p prevents recombination and SIR2 loss of function results in extrachromosomal rDNA circles (ERCs) accumulation.
Since Ubp10p is also associated with rDNA regions (Emre et al. 2005), we have investigated here if Ubp10 deubiquitinating activity could be involved in rDNA locus control by examining ERCs as a marker of rDNA recombination, regardless of their role in replicative senescence. All yeast strains used in this study are listed in Table 1. Genomic DNAs isolated from sir2, ubp10 null mutants and their isogenic wild-type strain were analyzed by two-dimensional (2D) chloroquine gels and probed for rDNA sequences. Mobilities of both linear and nicked circular DNA are unaffected by chloroquine concentration and they migrate along the diagonal of the gel. Supercoiled DNA circles form arcs that lie off the diagonal with the highly negatively ones running in the lower region of the arc (Sinclair and Guarente 1997). As shown in Figure 1A, ubp10 disruptant cells displayed ERCs accumulation. As a control, the ERCs pattern obtained for a sir2Δ mutant was also shown; this pattern did not change appreciably in the double ubp10 sir2 null mutant.
FOB1-dependent replication block causes a DNA double-strand break within the rDNA and this break can be repaired by homologous recombination, resulting in the formation of ERCs. Fob1 mutants have a reduced rate of ERCs formation (Defossez et al. 1999). Deletion of FOB1 reduced ERCs levels in the ubp10 background below those detected in wild-type cells (Figure 1A). An analogous reduction was observed following FOB1 inactivation in the wild-type strain (Figure 1A) in agreement with published data (Defossez et al. 1999; Kaeberlein et al. 1999). The same results were obtained for sir2 fob1 and ubp10 sir2 fob1 null mutants (data not shown). Taken together these data indicate that the sole lack of Ubp10 histone-deubiquitinating activity is able to determine ERCs accumulation and that ERCs are generated by a mechanism depending upon blocked replication forks.
Each rDNA repeat contains an origin of replication that allows the excised DNA circles to behave like autonomously replicating plasmids without a centromeric sequence. A highly asymmetric segregation of ERCs at cell division leads to ERCs accumulation in aged mother cells and assures that daughters are born ERCs free (Sinclair and Guarente 1997). To examine whether UBP10 deletion gave rise to a premature excision of ERCs, we isolated ubp10 mutant and wild-type cells of different replicative ages. Since the increase in size is a defining distinction between young and old cells, both strains were grown for eight generations and then size selected by centrifugal elutriation (Bitterman et al. 2003). Different fractions were collected and characterized by analyzing cell volume distributions (Figure 1B), by counting the number of bud scars after Calcofluor staining, and by determining the percentage of budded cells. In particular, the cell volume distributions of the first and the last elutriated fractions displayed quite separate profiles with different shapes (Figure 1B), in agreement with the presence of two types of cellular populations. Fractions 1 contained uniform populations of small unbudded daughter cells (>90% with no bud scars) with an average cell size of ∼28 fl and 25 fl for the wild-type and ubp10 mutant strains, respectively. Fractions 10 were enriched in large mother cells carrying six to eight bud scars whose broadened cell volume distributions had an average value of 97 fl for the wild-type and 95 fl for ubp10 cells. One-dimensional gel analyses were then performed on DNAs isolated from the different fractions to analyze the presence of ERCs. As shown in Figure 1C, in addition to a strong signal from the genomic rDNA, the rDNA probe detected two ERC species, monomers and dimers; the latter displayed a double band probably due to torsional differences as previously observed (Takeuchi et al. 2003). In young wild-type cells, a very small amount of ERCs was visible; the amount gradually increased along with the size increase. In the ubp10 strain, ERCs levels, as well as the rate of their increase, were higher (Figure 1C), indicating that UBP10 loss of function affects rDNA locus control similarly to SIR2 loss of function.
Suppression of recombination occurs through the establishment of a repressive/nonaccessible/silenced structure requiring Sir2 histone-deacetylase activity. Moreover, Ubp10p is required for optimal binding of Sir proteins to telomeres and global telomeric silencing (Orlandi et al. 2004; Emre et al. 2005; Gardner et al. 2005). This finding raised the possibility that Ubp10p might influence Sir2p association with the rDNA locus. Therefore, we generated wild-type and ubp10 strains in which the endogenous copy of SIR2 was epitope tagged at the C terminus, using the 3HA-KlURA3 module (Longtine et al. 1998). Tagged Sir2p was fully functional (see supplemental Figures 1 and 2 at http://www.genetics.org/supplemental/) and it showed comparable total cellular levels in both strains (Figure 2A). To analyze Sir2p distribution, chromatin immunoprecipitation (ChIP) experiments were performed with anti-HA antibodies. Immunoprecipitates (IP), as well as the corresponding whole-cell extracts (input), from each strain were assayed for coprecipitated DNA by PCR with primer pairs that amplify fragments, indicated in Figure 2B, spanning two preferential localization sites of Sir2p within nontranscribed spacer 1 (NTS1) and NTS2 regions of an rDNA repeat (Gotta et al. 1997; Huang and Moazed 2003). A 265-bp fragment located 52 bp from the start of the TG1-3/CA1-3 tract on the right telomere of chromosome VI (TEL VIR) and a 372-bp fragment within ARO1, a nontelomeric gene, were also amplified and used as positive and normalizing controls, respectively. As shown in Figure 2C, in wild-type cells tagged Sir2p was associated with TEL VIR and with both NTS1 and NTS2 regions (Gotta et al. 1997; Strahl-Bolsinger et al. 1997; Suka et al. 2002; Huang and Moazed 2003; Emre et al. 2005). UBP10 deletion affected Sir2p presence not only at the telomere, in agreement with the changes observed by (Emre et al. 2005), but also at the rDNA (Figure 2C). In fact, in the ubp10 strain the amount of amplified PCR products corresponding to NTS1 and NTS2 was reduced, showing, on average, a 1.8- and a 2.8-fold decrease, respectively (Figure 2D).
NTS1 and NTS2 are arranged in nucleosomal structures that display hypoacetylation of histone tails in a Sir2p-dependent manner (Armstrong et al. 2002; Bryk et al. 2002; Buck et al. 2002; Hoppe et al. 2002; Huang and Moazed 2003). Given that UBP10 deletion causes a decrease in Sir2p level at the NTS regions, we wondered whether this would correlate to an increase in histone acetylation, as well. In ubp10Δ, sir2Δ, and wild-type strains, ChIP experiments were performed as described above using antibodies that recognize general acetylation of histone H4 tails and the acetylated form of Lys16 of H4 (the target residue of Sir2 deacetylase activity). In addition to the expected results for SIR2 deletion, Figure 3, A and B, shows that UBP10 deletion also increased H4 acetylation at telomere and NTS regions as a likely consequence of the reduction in Sir2p level determined by ChIP (Figure 2). Finally, as a further refinement of our study, we measured Lys4 and Lys79 trimethylation of histone H3. ChIP analyses revealed that the levels of both modifications increased at the two NTS regions in the ubp10Δ mutant compared to those in wild-type cells (Figure 3, C and D). Measurements of H3 Lys4 and Lys79 trimethylation at TEL VIR were also performed as a positive control (Figure 3, C and D) and showed a degree of enrichment in ubp10 cells similar to previously reported data (Emre et al. 2005; Gardner et al. 2005). ChIP analyses performed in HA-tagged strains gave similar results (data not shown).
To obtain rDNA silencing, different cellular factors collaborate or compete, leading to an rDNA chromatin structure that is repressive to transcription of a RNA polymerase II reporter gene and recombination (Rusche et al. 2003; Machin et al. 2004; Mueller et al. 2006). We show here that the deubiquitinating enzyme Ubp10 contributes to the formation of such a structure by affecting Sir2p association. rDNA chromatin is highly responsive to alterations in SIR2 dosage and NTS regions represent the major location of SIR2-dependent alterations that have been detected in the rDNA array (Fritze et al. 1997; Smith et al. 1998; Cioci et al. 2002). UBP10 loss of function results in a decrease of Sir2p level at both NTS1 and NTS2 that correlates with histone hyperacetylation and, consequently, produces a more open chromatin configuration. Notably, the NTS1 fragment analyzed overlaps the replication fork block (Kobayashi 2003), suggesting that accumulation of ERCs observed in the ubp10Δ mutant can be ascribed to the reduced extent of Sir2p-dependent silent chromatin required to counteract Fob1p-dependent rDNA recombination at this region.
A Ubp10p requirement for a proper Sir2p localization at the rDNA is consistent with the enrichment of this deubiquitinating enzyme at the locus where it maintains low histone H3 trimethylation (Emre et al. 2005 and this work). Moreover, recalling that Sir4p targets Ubp10p at the telomeres to deubiquitinate H2B by optimizing association of Sir proteins (Gardner et al. 2005), Ubp10p could maintain a proper state of histone modification at the rDNA necessary for Sir2 binding. Clearly, further experiments are needed to determine additional partners involved in Ubp10p recruitment to the rDNA.
This work was supported by grants from Fondo per gli Investimenti della Ricerca di Base to L.A. and Fondo d'Ateneo per la Ricerca 2005 to M.V.
Communicating editor: M. Hampsey
- Received July 7, 2006.
- Accepted September 12, 2006.
- Copyright © 2006 by the Genetics Society of America