The origin recognition complex interacts with a subset of metabolic genes tightly linked to origins of replication

Abstract

The origin recognition complex (ORC) marks chromosomal sites as replication origins and is essential for replication initiation. In yeast, ORC also binds to DNA elements called silencers, where its primary function is to recruit silent information regulator (SIR) proteins to establish transcriptional silencing. Indeed, silencers function poorly as chromosomal origins. Several genetic, molecular, and biochemical studies of HMR-E have led to a model proposing that when ORC becomes limiting in the cell (such as in the orc2-1 mutant) only sites that bind ORC tightly (such as HMR-E) remain fully occupied by ORC, while lower affinity sites, including many origins, lose ORC occupancy. Since HMR-E possessed a unique non-replication function, we reasoned that other tight sites might reveal novel functions for ORC on chromosomes. Therefore, we comprehensively determined ORC "affinity" genome-wide by performing an ORC ChIP-on-chip in ORC2 and orc2-1 strains. Here we describe a novel group of orc2-1-resistant ORC-interacting chromosomal sites (ORF-ORC sites) that did not function as replication origins or silencers. Instead, ORF-ORC sites were comprised of protein-coding regions of highly transcribed metabolic genes. In contrast to the ORC-silencer paradigm, transcriptional activation promoted ORC association with these genes. Remarkably, ORF-ORC genes were enriched in proximity to origins of replication and, in several instances, were transcriptionally regulated by these origins. Taken together, these results suggest a surprising connection among ORC, replication origins, and cellular metabolism.

Figure 1. Resistance to orc2-1 defined a novel class of ORC binding sites.

(A) ORC binding to HMR-E in vivo is orc2-1-resistant. ORC ChIPs were performed in the ORC2 and orc2-1 strains at the permissive temperature of 23°C, and relative enrichment of HMR-E- or ARS1-containing DNA fragments was measured by PCR. ADH4 signal was used as a measure of background ORC binding. (B) The majority of sites identified in our ORC ChIP–on–chip have already been annotated by the Origin Database, validating our identification of bona fide ORC sites. (C) Novel ORC sites form an orc2-1-resistant, ORF–overlapping cluster. Each ORC peak that overlapped at least one ORF was included in the calculation outlined the diagram. D_max was calculated as described in the text and plotted against orc2-1/WT ratio. ORC peaks in the graph are classified according to their OriDB annotations. “Novel” sites are those identified for the first time in this study as ORC–binding sites.

Figure 2. Detection of ORF–ORC sites in directed ChIP experiments.

(A) Examples, from the ChIP–on–chip data, of an orc2-1-sensitive ORC peak at a replication origin, ARS820, and of a nearby orc2-1-resistant ORF–ORC peak at the ENO2 gene. (B) Immunoprecipitation with a cocktail of monoclonal antibodies against Orc1p, Orc2p, and Orc3p resulted in a slight enrichment of the ENO2 ORF relative to a background locus (FKH1), while immunoprecipitation with a cocktail of two monoclonal Sir3 antibodies did not result in any enrichment of ENO2. On the other hand, the Sir3 antibodies, similarly to the ORC antibodies, were able to efficiently immunoprecipitate HMR. (C) ChIPs with an anti–HA antibody were performed in Orc2-3xHA and untagged strains. Association of the antibody with two different ORF-ORC loci, ENO2 and TDH3, as well as an origin of replication, ARS820, was dependent on the presence of HA–tagged Orc2. All data were normalized to immunoprecipitated FKH1 from the same sample. Averages of two biological replicates of each strain are plotted, with error bars representing one standard deviation.

Figure 3. ORF–ORC sites did not function as origins of replication on the chromosome.

The diagram shows expected migration patterns of different replication intermediates (replication bubbles and forks) on 2D gels. The ratio of bubbles to small forks corresponds to origin efficiency. Restriction enzyme and probe positions (black rectangles) are shown. The two “confirmed” origins, ARS731.5 and ARS820, are orc2-1-sensitive (with orc2-1/WT ratios of 0.36 and 0.35, respectively), corresponding to a decrease in firing at these sites in the orc2-1 mutant. Both ORF–ORC sites, TDH3 and ENO2, have been detected as sites of ORC and MCM binding in another ChIP–based study and are annotated as “likely” ARSs by OriDB. In the figure, dashed ovals indicate “likely” ARS boundaries. Both “likely” ARSs at TDH3 and ENO2 are orc2-1-resistant (orc2-1/WT ratios of 1.1 for each), but neither one produced replication bubbles either in the wild-type or in the orc2-1 strain. Gray bars indicate chromosomal clones tested in plasmid origin assays (Table 1, Table S3).

Figure 4. Origins of replication could combine orc2-1-resistance with high firing efficiency.

Unlike HMR-E, orc2-1-resistant replication origin ARS1005 could fire with high efficiency at its endogenous chromosomal location. Like HMR-E, ARS1005 was resistant to orc2-1 for firing efficiency, while neighboring origin ARS1006 that was orc2-1-sensitive for ORC binding was also orc2-1–sensitive for firing.

Figure 5. Transcriptional activity promoted ORC binding to an ORF.

(A) Histograms of expression levels of the ORF–ORC set and of all genomic ORFs are shown. Dubious ORFs were not included in this analysis. (B) TDH3 promoter was replaced by the GAL1 promoter (pGAL1). pGAL1-TDH3 and wild-type strains were grown in glucose or galactose, RNA was isolated, and gene expression measured by reverse transcriptase PCR. Growth in glucose repressed TDH3 transcription in the pGAL1-TDH3 strain but not in an isogenic wild-type strain, whereas wild-type TDH3 was virtually unaffected by change of carbon source. As expected, expression of the GAL1 gene was repressed in glucose and induced in galactose. ENO2 expression was monitored as a control. (C) Results of directed ORC ChIPs on the wild-type and pGAL1-TDH3 strains grown in glucose and galactose are shown. Averages of two to four independent biological replicates are plotted for each condition, with error bars representing one standard deviation. Repression of pGAL1-TDH3 transcription by glucose reduced ORC binding to TDH3 ORF, while induction of GAL1 in galactose increased ORC binding there by 2.5-fold.

Figure 6. ORF–ORC sites are enriched downstream of replication origins.

The schematic at the top of the figure shows how distances to the nearest 5′ and 3′ ARSs were calculated for every ORF in the genome (dubious ORFs were omitted from this analysis). (A) The ORF–ORC set was enriched for genes located within 10 Kb of the nearest upstream origin compared to all ORFs. (B) Median distances to nearest 5′ and 3′ origins are plotted for several classes of ORFs, categorized by expression level or functional process. The ORF–ORC set was the only gene category closely associated with upstream origins.

Figure 7. Deletion of a replication origin reduced expression of a downstream ORF–ORC.

(A) Repressing TDH3 transcription by growing the pGAL1-TDH3 strain in glucose did not grossly alter firing efficiency of the nearby replication origin ARS731.5 compared to a wild-type strain with a highly transcribed TDH3 gene. (B) ORC–binding sites within three different replication origins were deleted separately, and both ORC association and gene transcription within the three regions were analyzed. The diagrams show relative positions of genes around the origins. The ORF–ORC genes are shown as black arrows with white font. In each case, deleting the origin's ORC binding site abolished ORC association with the origin but not ORC association with the nearby ORF–ORC. Also in each case, deletion of a replication origin reduced transcription of the nearby ORF–ORC gene while leaving other surrounding genes relatively unaffected. The averages of at least two biological replicates are plotted on the graph, with error bars representing one standard deviation. For the experiment in first panel (ARS731.5-TDH3 region), quantification of expression of genes near ARS731.5 was done using band densitometry. For panels 2 and 3, quantification of expression of genes near ARS820 and ARS1627 was done using real-time PCR (Materials and Methods).