Roles for Gcn5 in promoting nucleosome assembly and maintaining genome integrity

Abstract

The coordinated process of DNA replication and nucleosome assembly, termed replication-coupled (RC) nucleosome assembly, is important for the maintenance of genome integrity.  Loss of genome integrity is linked to aging and cancer. RC nucleosome assembly involves deposition of histone H3-H4 by the histone chaperones CAF-1, Rtt106 and Asf1 onto newly-replicated DNA.  Coordinated actions of these three histone chaperones are regulated by modifications on the histone proteins.  One such modification is histone H3 lysine 56 acetylation (H3K56Ac), a mark of newly-synthesized histone H3 that regulates the interaction between H3-H4 and the histone chaperones CAF-1 and Rtt106 following DNA replication and DNA repair.  Recently, we have shown that the lysine acetyltransferase Gcn5 and H3 N-terminal tail lysine acetylation also regulates the interaction between H3-H4 and CAF-1 to promote the deposition of newly-synthesized histones. Genetic studies indicate that Gcn5 and Rtt109, the H3K56Ac lysine acetyltransferase, function in parallel to maintain genome stability. Utilizing synthetic genetic array analysis, we set out to identify additional genes that function in parallel with Gcn5 in response to DNA damage. We summarize here the role of Gcn5 in nucleosome assembly and suggest that Gcn5 impacts genome integrity via multiple mechanisms, including nucleosome assembly.

Nucleosome Assembly Following DNA Replication and DNA Repair

Chromatin structure plays a significant role in several cellular processes including DNA replication, transcription and DNA repair.1,2 The fundamental unit of chromatin is the nucleosome which consists of 147 base pairs of DNA wrapped around the histone octamer containing two copies each of the core histones H2A, H2B, H3, and H4. During DNA replication, nucleosomes ahead of the replication fork are disassembled to allow DNA replication machinery access to the DNA. Immediately following DNA synthesis, replicated DNA must be assembled into nucleosomes and subsequent higher order chromatin structure. This process of coordinated DNA replication and nucleosome assembly, referred to as DNA replication-coupled (RC) nucleosome assembly,3,4 is important for epigenetic inheritance and the maintenance of genome integrity.1,2,5 Loss of genome integrity is one of the key characteristics of cancer cells and aging.6,7 Therefore, understanding how RC nucleosome assembly is regulated is extremely important.

There are two pools of histones used for formation of nucleosomes following DNA replication: parental and newly-synthesized histones. While it is clear that assembly of newly-synthesized H3–H4 requires histone chaperones, it is not clear how parental histones are assembled. Histone chaperones are key proteins involved in the regulation of nucleosome assembly. By binding to histones, these proteins help shield the positively-charged his-tones from the negatively-charged DNA and facilitate histone deposition as well as histone removal.2 In the yeast S. cerevisiae, the H3–H4 histone chaperones include Asf1, CAF-1 and Rtt106. These three his-tone chaperones function coordinately to efficiently deposit newly-synthesized his-tones onto the replicating DNA.2,8

Following repair of damaged DNA, nucleosomes must be reassembled and chromatin structure restored. This process is proposed to utilize similar machinery as replication-coupled nucleosome assembly. The current model to explain the role of chromatin structure and chromatin-associated proteins following DNA repair is the access-repair-restore model.9,10 Following DNA damage, repair factors must gain access to the DNA. Chromatin remodeling factors are involved in remodeling and/or disassembly of nucleosomes. Following repair of the lesion, nucleosomes and the chromatin structure must be reestablished. The histone chaperone CAF-1 is important for repair following DNA damage in both yeast and mammalian cells.11–13 In addition, the histone chaperone Asf1 is also linked to checkpoint activation associated with DNA damage through an interaction with Rad53.14 Given that many of the same factors are involved, it is probable that many of the same regulatory mechanisms also control nucleosome assembly following DNA repair.

Multiple Acetylation Events Have Been Proposed to Regulate Nucleosome Assembly.

Newly-synthesized histones are associated with particular post-translational modifications that are thought to aid in the recognition and deposition of these histones at the replication fork.2 Lysine acetylation is the most well-studied histone modification involved in the regulation of nucleosome assembly. Acetylation occurs prior to deposition, and the histones are deacetylated following deposition.15 It has been known for a long time that the most highly conserved marks of newly-synthesized his-tones are acetylation of histone H4 lysine residues 5 and 12 (H4K5,12Ac).16 These marks are catalyzed by Hat1.17,18 While this pattern is found on H4 in complex with CAF-1, drosophila and mammalian Asf1, and yeast Hif1,17,19–21 the precise role of this mark in nucleosome assembly remains to be determined.

In yeast, newly-synthesized histone H3 is marked by lysine 56 acetylation (H3K56Ac).22 This mark, catalyzed by the lysine acetyltransferase Rtt109, is found within the core domain of H3 and is dependent on the histone chaperone Asf1.23–26 H3K56Ac mediates the interaction between H3-H4 and CAF-1 and Rtt106, thereby promoting replication-coupled nucleosome assembly.27 In yeast, it has been found that H3K56Ac helps drive chromatin assembly following repair and signals the completion of repair.28 In mammalian cells, H3K56Ac has been found to have a role in DNA damage and repair. In addition, H3K56Ac has been associated with tumorigenicity, consistent with a role in genomic stability.29,30 The histone acetyltransferases responsible for H3K56ac in mammalian cells include CBP/p300 and Gcn5.29,31 Further studies are needed to fully understand the function of H3K56Ac in mammalian cells.

In addition to H3K56Ac, acetylation of H3 N-terminal tails has also been reported. However, patterns of acetylation of the N-terminus of newly-synthesized H3 are not as well conserved.16 For instance, in yeast cells, newly-synthesized H3 is acetylated predominantly at lysine 9 followed by lysine 27.16,32 In mammalian cells, acetylation of H3.1, the H3 variant incorporated into chromatin during DNA replication, is barely detectable.16 However, a recent study found that H3 lysine 14 and 18 are present on 20–30% of H3.1 in association with Asf1 during S-phase in HeLa cells.21 Given the low conservation among species, the role of acetylated lysine residues at the N-terminus of H3 in nucleosome assembly is not well studied. Furthermore, the lysine acetyltransferase that acetylates the H3-N-terminus and functions in nucleosome assembly, until recently, was not known.

Recent studies from our laboratory suggest that the histone acetyltransferases Elp3 and Gcn5 are involved in marking newly-synthesized histones and regulating nucleosome assembly.33,34 Here, we review the role of Gcn5 in replication-coupled nucleosome assembly following DNA replication. In addition, we report the results of a synthetic genetic array (SGA) screen performed to identify genes that function in parallel to Gcn5 in response to DNA damage. We found that multiple pathways function in parallel with Gcn5 in the maintenance of genome integrity.

A Role for the Lysine Acetyltransferase Gcn5 in Replication-Coupled Nucleosome Assembly

Gcn5 (general control non-derepressible 5) was originally identified in a screen for factors involved in the regulation of amino acid biogenesis.35,36 Later studies showed that Gcn5 is a transcription activator37 and has lysine acetyltransferase activity and that this activity contributes towards its role in transcriptional activation.38,39 Gcn5 is highly conserved and is a catalytic component of several large complexes within yeast and higher eukaryotes.40–42 Gcn5 by itself has limited histone lysine acetyltransferase activity, namely acetylation of histone H3K14 as well as H4 K8 and 16 of free histones.32 This activity is greatly expanded when Gcn5 is part of the larger ADA and SAGA complexes, which enable acetylation of nucleosomal histones.43 Besides their shared components, both the ADA and SAGA complexes regulate gene transcription, in part, through other components of these complexes.42,44 For instance, Spt8, a unique component of the SAGA complex is involved in recruitment of TBP to active promoters, as is the component Spt3.45,46 In addition, the Sus1-Sgf11-Ubp8 ubiquitin complex of SAGA is involved in H2B deubiquitination as well as regulating the level of H3K4me3.47–50 While Gcn5 and its complex members' roles in transcriptional activation are well established, it has recently become clear that Gcn5 has roles beyond transcriptional regulation.

Cells that lack Gcn5 are sensitive to DNA damaging agents and exhibit G₂/M arrest, suggesting that Gcn5 has a role in the maintenance of genome integrity.51 While Gcn5′s role in transcription may help explain some of these phenotypes, it certainly doesn't explain them completely, therefore, suggesting that Gcn5 and histone N-terminal tail acetylation functions in processes besides gene transcription.52 Our laboratory and others found that the histone H3K56 acetyltransferase, RTT109, exhibits a significant genetic interaction with GCN5 and H3 N-terminal tail acetylation in growth and DNA damage sensitivity, indicating that Gcn5 and H3 N-terminal tail acetylation function in parallel with H3K56Ac.27,34,53 Compared to wild-type cells, both rtt109Δ and gcn5Δ mutant cells exhibit a higher percentage of Rad52 foci, a measure of spontaneous DNA damage. In addition, GCN5 exhibits a synthetic genetic interaction with genes encoding checkpoint kinases (RAD53 and MEC1), DNA replication proteins (CDC17 and CDC7), and histone chaperones (CAF-1, RTT106, and ASF1). These observations suggest that Gcn5′s role in the maintenance of genome integrity may lie, at least partly, in a role in DNA replication-coupled nucleosome assembly.34 Supporting this idea, we have shown that the deposition of newly-synthesized H3, marked by H3K56Ac, is compromised in cells lacking Gcn5 or mutant cells unable to acetylate the N-terminus of H3. Furthermore, Gcn5 and N-terminal tail acetylation promote the interaction between H3-H4 and CAF-1.34 These results provide strong support for the idea that Gcn5 has a role in replication-coupled nucleosome assembly, in part, through the acetylation of lysine residues at the H3 N-terminus.

Interestingly, while genetic evidence suggests that GCN5 and RTT109 function in parallel in response to DNA damage, our studies suggest a crosstalk between Gcn5 and Rtt109 during replication-coupled nucleosome assembly. H3K56Ac, catalyzed by Rtt109, increases the binding of H3–H4 with both Rtt106 and CAF-1.27 In contrast, Gcn5 and acetylation of lysine residues at the H3 N-terminus impact the binding of CAF-1 to H3–H4 and have little impact on the association of Rtt106 with H3–H4. Moreover, Rtt109 can also acetylate lysine 9 and lysine 27 of H3 N-terminus, two residues that can also be acetylated by Gcn5.34,54 However, it is important to note that Gcn5, but not Rtt109, is the predominant acetyltransferase for these two lysine residues. In addition, while it is reported that Gcn5 can acetylate H3 K56 in mammalian cells,31 in budding yeast, deletion of GCN5 has no apparent effect on H3K56Ac in cells.34 These results suggest that although there is crosstalk between the two enzymes, they predominantly function in parallel pathways to regulate nucleosome assembly and maintain genome stability. Thus, Gcn5 and Rtt109 function together to promote nucleosome assembly and genome stability (Fig. 1A).

Roles for Gcn5 in the Initiation of DNA Replication and Repair

Mutant gcn5Δ cells are significantly more sensitive to the DNA damaging agents hydroxyurea (HU), camptothecin (CPT), and methyl methanesulfonate (MMS) than cells lacking CAC1, the large subunit of CAF-1.34 This suggests that Gcn5 (and furthermore, H3 N-terminal tail acetylation) may have functions beyond its role in regulating the interaction between CAF-1 and H3–H4 that contribute towards its general function in response to DNA damaging agents.34 Our lab and others have found that Gcn5 also directly associates with origins of DNA replication. Moreover, cells lacking Gcn5 have a deficiency in the assembly of pre-replicative complexes and a significant disruption of the chromatin structure surrounding the replication origin.55 These results suggest that Gcn5 has a role in the initiation of DNA replication.34,55 It is thought that Gcn5 may serve to decondense the chromatin structure to facilitate access for DNA replication machinery.55 Supporting this idea, it has been found that histone N-terminal tail acetylation positively regulates origin firing in yeast cells55,56 and that acetylation of multiple sites on H3 and H4 regulates origin firing.56 Interestingly, acetylation of many of these same sites has been implicated or shown to have roles in replication-coupled nucleosome assembly. Because defects in nucleosome assembly can impact the integrity of chromatin structure, the observed defects in the initiation of DNA replication in cells lacking Gcn5 could be partly due to an indirect effect arising from defects in nucleosome assembly.

In mammalian cells, it's been found that Gcn5 acetylates Cdc6, a component of the pre-replication complex, stimulating its phosphorylation and activation.57 While this interaction has not yet been observed in yeast, it is still possible that Gcn5 also contributes to origin firing by acetylating proteins involved in the initiation of DNA replication. Future studies are needed to address how Gcn5 regulates the initiation of DNA replication.

Gcn5 is also implicated in DNA repair. In yeast, Gcn5 is recruited to double strand break sites, and gcn5Δ cells are sensitive to DNA double strand breaks. In conjunction with a role for Gcn5 in double strand break repair, acetylation of multiple sites of the H3 N-terminus increases following a double strand break.58 In addition, it has been shown that acetylation of H3 and Hat1, which catalyzes acetylation of lysine 5 and 12 of H4, is important for double strand break repair.59 In mammalian cells, the STAGA and TFTC complexes, two of the large Gcn5-containing complexes in mammalian cells, are involved in UV damage repair.60,61 Despite these advances, it is still unclear how Gcn5 and acetylation of lysine residues at the H3-N-terminus function in DNA repair.

Given that Gcn5 and Rtt109 function in parallel in the response to DNA damaging agents, and Rtt109 promotes nucleosome assembly following both DNA replication and DNA repair,27,28,34 it is possible that Gcn5 and acetylation of lysine residues of the H3 N-terminus function in DNA repair by promoting assembly of replicated DNA into nucleosomes. More recently, it has been shown that H3 N-terminal tail acetylation by Gcn5 following DNA damage may aid in the recruitment of the chromatin remodeling SWI/SNF complex, facilitating DNA damage signaling and repair.62 Thus, Gcn5 and acetylation of lysine residues at the H3 N-terminus may impact the repair process through multiple means.

Identification of Genes that Function in Parallel with GCN5 to Maintain Genome Stability

Given Gcn5′s recently described roles in DNA replication and repair, we wished to identify other pathways that function in parallel with Gcn5 in the maintenance of genome integrity. Therefore, we set out to identify gcn5Δ double mutants that did not exhibit synthetic growth defects but were sensitive to the DNA damaging agent camptothecin (CPT). The drug CPT results in stabilization of the topoisomerase I-DNA intermediate, ultimately resulting in single-strand and double-strand breaks.63 We performed a yeast genetic screen using the synthetic genetic array (SGA) based strategy,64,65 combining the gcn5Δ mutant with each of ∼4700 yeast deletion mutants. We then challenged the viable double mutants with different concentrations of CPT (Fig. 2A). Double mutant cells were scored for growth defects and CPT sensitivity. These mutant cells were verified by individual spot assays where cells of a given genotype were plated in a ten-fold series dilution onto media containing low concentrations of CPT. Finally, we deleted each candidate gene in our laboratory background, W303, and tested how the cells responded to CPT alone or in combination with the gcn5Δ mutant (Fig. 2B and supplemental figure S1). From these tests, we identified ten mutants that were more sensitive to CPT when combined with gcn5Δ than either single mutant alone (Fig. 2C). Verifying our screening strategy, one of the genes identified was RTT109, a gene we have shown to function in parallel with GCN5 in the maintenance of genome stability. In addition to CPT, we also determined how each double mutant responded to the DNA damaging agents HU and MMS (Fig. 2B, S2 and Table 1). We found that most of the double mutants were sensitive to all DNA damaging agents tested, suggesting that GCN5 and each of these genes are involved in the DNA damage stress response.

The genes that exhibited a synthetic phenotype with GCN5 in response to CPT could be classified into the following groups: checkpoint activation, DNA repair, and other pathways. First, we found that GCN5 genetically interacts with genes involved in checkpoint activation such as RAD9 and RAD24.66 This result is consistent with our published findings that GCN5 genetically interacts with the checkpoint kinases RAD53 and MEC1 and that gcn5Δ mutant cells experience spontaneous chromosome breaks.34 Interestingly, GCN5 only exhibited a synthetic phenotype with RAD9 and RAD24 in response to CPT and HU, but not MMS. Furthermore, the synthetic interaction observed in these double mutants was not as striking as the interaction observed between other genes identified, such as RTT109 (Table 1).

Second, we observed that GCN5 genetically interacted with genes implicated directly and indirectly in DNA repair. These genes include MMS4, RPN4, DCC1, and YTA7. Mms4 is a subunit of the Mms4-Mms81 endonuclease involved in DNA damage repair and homologous recombination.67,68 Rpn4 is transcription factor that regulates expression of proteasome genes. In addition, Rpn4 is regulated by various forms of cellular stress69 and localizes to double strand break sites.70 Dcc1 forms a complex with Ctf18 and Ctf8 and is essential for the establishment of sister-chromatid cohesion and DNA replication.71 Sister-chromatid cohesion proteins, including Dcc1, have been found to be important for DNA damage repair.72 Yta7 is predicted to control histone gene expression and occupies similar regions as the histone chaperone Rtt106, Spt4/5/6, and the FACT complex.73 In addition, Yta7 has been identified as one of many proteins in association with chromatin near replication origins.56 It's been suggested that Yta7 functions in an overlapping pathway with Asf1, Hir1 and FACT in response to stress, such as DNA damage.74 This is consistent with our findings that gcn5Δ yta7Δ cells do not show growth defects but do have significant DNA damage sensitivity. It is possible that both Gcn5 and Yta7 control gene expression during times of cellular stress, but it could also suggest that these proteins play a more direct role in DNA damage repair and genome integrity.

Finally, we also observed a significant synthetic interaction between GCN5 and three other genes, MDM20, RPL14A and ELM1. Deletion of each of these three genes, when combined with the gcn5Δ mutant, showed significant sensitivity to all three DNA damaging agents tested (Table 1). MDM20 is a subunit of the NatB N-terminal acetyltransferase complex.75 Interestingly, Cac2, one of the components of the CAF-1 complex, is a substrate of NatB. Other potential substrates of the NatB complex include proteins involved in nucleosome assembly, cell cycle progression and DNA processing.76 Rpl14A is a component of the 60S ribosome,77 whereas Elm1 is a serine/threonine kinase with a role in the regulation of cytokinesis78 and determination of cellular morphology.79 While the functions of these three proteins in response to DNA damage stress are unclear, our studies suggest a role for these gene products in the maintenance of genome integrity.

In summary, we found that GCN5 exhibits synthetic interactions with several genes involved in DNA damage response and DNA repair. These genetic interactions provide additional evidence that Gcn5 has a role in the maintenance of genome integrity, and this role may be independent of its role in gene transcription.

Overall Summary

Here, we reviewed recent studies indicating that Gcn5 may have a multitude of functions contributing towards the maintenance of genome stability including roles in DNA replication-coupled nucleosome assembly, initiation of DNA replication, and DNA repair. We also described a screen we performed which identified ten genes that function in parallel to GCN5 in response to DNA damage stress. We hypothesize that one of the primary functions of Gcn5 in DNA replication and repair is the regulation of nucleosome assembly. Gcn5's role in nucleosome assembly may explain many of the phenotypes observed in gcn5Δ cells such as the defects in origin firing and sensitivity to DNA damage agents. Moreover, because cells defective in nucleosome assembly, such as cells lacking CAF-1, are known to affect gene transcription globally,80 it would be interesting to determine to what extent the transcriptional defects observed in gcn5Δ mutant cells are due to compromised nucleosome assembly in this mutant cells. Future studies will shed light on molecular mechanisms by which acetylation and the modifying enzymes including Gcn5 are involved in the processes of DNA replication and repair, two processes contributing to the maintenance of genome integrity.

Figures and Tables

Figure 1 A role for Gcn5 in replication-coupled nucleosome assembly. Rtt109 and Gcn5 are two acetyltransferases that acetylate newly synthesized histone H3. H3 acetylation increases the binding of histones with histone chaperone proteins (CAF-1 and Rtt106) and promotes the deposition of histones onto the newly replicated DNA. Rtt109 predominantly acetylates H3K56, which is important for both CAF-1 and Rtt106 to bind H3. Gcn5 acetylates lysine residues of the H3 N-terminal tail, including H3 lysines 9, 14, 18, 23, and 27, which is important for CAF-1 to bind H3. In addition, Rtt109 is able to acetylate additional residues at the histone H3 N-terminal tail, and this acetylation is a minor contribution compared to Gcn5 (as depicted via a dashed arrow).

Figure 2 A genetic screen was performed to identify genetic pathways that function in parallel to GCN5 in response to DNA damaging agent CPT. (A) Schematic diagram of the yeast genetic screen for mutants that enhance the sensitivity of gcn5Δ mutant to the DNA damaging agent, camptothecin (CPT). (B) Mutants from the initial screen were evaluated using a spot assay. A ten fold series dilution of yeast cells of the indicated genotype were plated onto regular growth media, YPD, or media containing low concentrations of the DNA damage agents camptothecin (CPT), methyl methanesulfonate (MMS), and hydroxyurea (HU). (C) Summary of the genes found to interact genetically in parallel with GCN5 in response to DNA damage. Those genes in red have roles in checkpoint activation, those in blue have identified roles in DNA repair, and those in green are genes with other functions that act in a parallel manner to GCN5 in response to DNA damage.

Table 1 Genes that exhibit synthetic genetic interactions with GCN5 in response to DNA damaging agents

Gene Name	Growth	CPT	HU	MMS
RTT109	+++	++++	++++	++++
DCC1	−	+++	+++	+++
MDM20	−	+++	+++	++
RPL14A	−	+	++	−
ELM1	+++	++	++++	+
RAD9	−	+	+	+
MMS4	−	++++	+++	++
RAD24	−	+	+	−
RPN4	−	+++	++++	−
YTA7	−	++++	+++	+

Genetic interactions between the gcn5Δ mutant and the indicated gene deletion mutant in growth and DNA damage sensitivity to CPT, HU, and MMS. Growth and DNA damage sensitivity was assessed via spot assays described in Figure 2B. + represents a synthetic effect in growth or sensitivity to the indicated DNA damaging agent. Additional + indicates that double mutant cells had more significant growth or DNA damage sensitivity (slower growth in response to the indicated challenge).

Acknowledgements

R.J.B. is supported by an American Heart Association Predoctoral Fellowship and the Mayo Clinic Sydney Luckman Family Predoctoral Fellowship. Z.Z. is a scholar of the Leukemain and Lymphoma Society. Work in the laboratory of Z.Z. is supported by grants from the National Institues of Health.