Genome-wide sequence and functional analysis of early replicating DNA in normal human fibroblasts

Abstract

Background: The replication of mammalian genomic DNA during the S phase is a highly coordinated process that occurs in a programmed manner. Recent studies have begun to elucidate the pattern of replication timing on a genomic scale. Using a combination of experimental and computational techniques, we identified a genome-wide set of the earliest replicating sequences. This was accomplished by first creating a cosmid library containing DNA enriched in sequences that replicate early in the S phase of normal human fibroblasts. Clone ends were then sequenced and aligned to the human genome.

Results: By clustering adjacent or overlapping early replicating clones, we identified 1759 "islands" averaging 100 kb in length, allowing us to perform the most detailed analysis to date of DNA characteristics and genes contained within early replicating DNA. Islands are enriched in open chromatin, transcription related elements, and Alu repetitive elements, with an underrepresentation of LINE elements. In addition, we found a paucity of LTR retroposons, DNA transposon sequences, and an enrichment in all classes of tandem repeats, except for dinucleotides.

Conclusion: An analysis of genes associated with islands revealed that nearly half of all genes in the WNT family, and a number of genes in the base excision repair pathway, including four of ten DNA glycosylases, were associated with island sequences. Also, we found an overrepresentation of members of apoptosis-associated genes in very early replicating sequences from both fibroblast and lymphoblastoid cells. These data suggest that there is a temporal pattern of replication for some functionally related genes.

Figure 1

Flow cytometry of synchronized NHF1 cells. Cells were grown to confluence arrest and replated at low density in the presence of aphidicolin and BrdUrd as described. After 24 hrs cells were collected and analyzed by flow cytometry. The ordinate of this graph shows BrdUrd incorporation as indicated by fluorescently-labeled anti-BrdUrd antibody and the abscissa shows DNA content as indicated by propidium iodide staining. The boxes outline areas that were analyzed for cell number.

Figure 2

Island represented in mirror of UCSC Genome Browser. The early replication island chr12.389 is displayed as an annotation track in a mirror copy of the UCSC Genome Browser and is available at [29]. The cosmid clones that support this island, ERC2-43A12 and ERC3-39D2, are also shown below. This island overlaps a BAC clone, RP11-302B13, for which both replication timing [18, 19] and an open chromatin status [22] are available. The dark boxes indicate that this region was found to replicate early (timing ratio = 2.00) and to have an open chromatin conformation (log2 open = 2.73). The 200-kb region displayed is gene dense, including two WNT genes, is highly enriched in SINE repeat elements, and shows a paucity of LINE, LTR, and DNA transposons.

Figure 3

Determination of the timing of replication of genetic markers by PCR. The three composite panels illustrate the procedure used for determining which of seven samples of DNA, each replicated at a different 1-h period of the S phase, was enriched for copies of a particular marker. DNA replicated at the indicated 1-h intervals of the S phase of synchronized fibroblasts was labeled with BrdUrd and isolated by CsCl centrifugation. In each panel, A: Inverted contrast image of PCR products stained with ethidium bromide after gel electrophoresis. The same primer set was used to amplify increasing amounts of genomic DNA and DNA replicated during each of the first 7 h of the S phase; a control with no template DNA was included in every PCR experiment. The R² value for each standard curve that was obtained by plotting the signal intensity of the bands of PCR products in the gel above versus the amount of genomic DNA, is shown. B: Bar graph illustrating the abundance of the marker in each of the seven 1-h samples of the S phase. Results show the average of two synchronizations that were each tested at least twice. Vertical bars indicate the standard deviation for each time point. Relative abundance was calculated from the linear regression equation of the standard curve and expressed as a percentage of the S phase fraction with the highest value. Replication times were calculated as described in the Methods section. The three panels shown illustrate the results obtained with markers that replicate in the first hour of the S phase (primer chr22.1096B), during the second hour (chr6.1385B), and at 4.6 hrs into S phase (chr6.1417A).