ES cells do not activate p53-dependent stress responses and undergo p53-independent apoptosis in response to DNA damage

BACKGROUND

Embryonic stem (ES) cells can contribute precursors to all adult cell lineages. Consequently, damage to ES cell genomes may cause serious developmental malfunctions. In somatic cells, cell-cycle checkpoints limit DNA damage by preventing DNA replication under conditions that may produce chromosomal aberrations. The tumor suppressor p53 is involved in such checkpoint controls and is also required to avoid a high rate of embryonic malformations. We characterized the cell-cycle and DNA-damage responses of ES cells to elucidate the mechanisms that prevent accumulation or transmission of damaged genomes during development.

RESULTS

ES cells derived from wild-type mice did not undergo cell-cycle arrest in response to DNA damage or nucleotide depletion, although they synthesized abundant quantities of p53. The p53 protein in ES cells was cytoplasmic and translocated inefficiently to the nucleus upon nucleotide depletion. Expression of high levels of active p53 from an adenovirus vector could not trigger cell cycle arrest. Instead, ES cells that sustained DNA damage underwent p53-independent apoptosis. The antimetabolite-induced p53-dependent arrest response was restored in ES cells upon differentiation.

CONCLUSIONS

Cell-cycle regulatory pathways in early embryos differ significantly from those in differentiated somatic cells. In undifferentiated ES cells, p53 checkpoint pathways are compromised by factors that affect the nuclear localization of p53 and by the loss of downstream factors that are necessary to induce cell-cycle arrest. A p53-independent programmed cell death pathway is effectively employed to prevent cells with damaged genomes from contributing to the developing organism. The p53-mediated checkpoint controls become important when differentiation occurs.

Mapping replication fork direction by leading strand analysis

Replication fork polarity methods measure the direction of DNA synthesis by taking advantage of the asymmetric nature of DNA replication. One procedure that has been used on a variety of cell lines from different metazoans relies on the isolation of newly replicated DNA strands in the presence of the protein synthesis inhibitor emetine. Since Okazaki fragments are not synthesized under such conditions, DNA strands produced during continuous exposure to emetine consist mainly of leading strands. In the protocol described, leading strands isolated from emetine treated cells are detected with single-stranded probes representing each strand of the DNA duplex in the region of interest. Hybridization of leading strands to one strand of a cloned genomic template identifies the direction of replication fork movement. If initiation of DNA synthesis occurs from a preferred site, leading strands should diverge from the corresponding initiation region. The leading strand method is particularly useful for mapping initiation in chromosomal loci that do not replicate immediately on entry into S phase and in mapping the replication fork patterns in which candidate initiation regions have not been identified. Cautious interpretation of the results is needed because the method relies heavily on quantitative hybridization. Leading strand data can be difficult to interpret when the genomic targets are very close to initiation regions and when the targets vary in their hybridization efficiency or in the efficiency of incorporation of nucleotide analogs. The experimental details of the method are reviewed, controls to avoid common pitfalls are suggested, protocols to facilitate the accurate interpretation of the results are provided.

Maintaining genetic stability through TP53 mediated checkpoint control

TP53 serves as a key relay for signals elicited by cellular stresses arising from diverse environmental or therapeutic insults. This relay then activates a cell cycle arrest or cell death program, depending on the stimulus and cell type. The absence of TP53 function disables the cell death or arrest programmes, thereby allowing the emergence of variants with various types of genomic alterations. The data discussed focus on two different types of signals that trigger the TP53 relay system. Firstly, TP53 arrests cell cycle progression in response to the types of DNA damage most commonly detected in cells undergoing tumour progression. Secondly, TP53 is activated by specific depletion of ribonucleotide pools, which prevent cells from entering S phase under conditions that could lead to chromosome breakage. The contribution of both responses limits the emergence of genetic variants. The DNA damage induced arrest appears to be triggered by as few as one double strand break in normal human fibroblasts. Analysis of the arrest kinetics after ionizing radiation shows that TP53 activates a prolonged arrest response in cells with irreparable DNA damage and that high efficiency cell elimination is achieved by a process that can be activated over multiple cell cycles. These data indicate that the primary function of the TP53 arrest/apoptosis pathway in response to double strand break is to eliminate damaged cells from the proliferating population, not to allow additional time for lesion repair. However, it remains possible that repair of other types of damage may benefit from TP53 mediated arrest. Analyses in model genetic systems indicate that the absence of TP53 function allows, but does not ensure, a high intrinsic rate of genetic variation and that instability is increased substantially when cells proceed through S phase under inappropriate growth conditions. This implies that inactivation of TP53 function in combination with other genetic alterations, such as oncogene activation, could accelerate genomic instability and tumour progression.

p53-mediated accumulation of hypophosphorylated pRb after the G1 restriction point fails to halt cell cycle progression

This study analyses whether the inability of p53 to induce G1 arrest after the restriction point relates to an inability to modulate pRb phosphorylation. Transient p53 overexpression in normal human diploid fibroblasts and p53-deficient cancer cells led to increased levels of the cyclin-dependent kinase inhibitor p21 cip1/Waf1/Sdi1 and an accumulation of hypophosphorylated pRb in cells growing asynchronously and in cells synchronized in late G1 or M. Similarly, gamma-irradiation of asynchronous, late-G1, or S phase fibroblasts led to an increase in hypophosphorylated pRb. Experiments with fibroblasts expressing the HPV16 E6 protein indicated that accumulation of hypophosphorylated pRb required functional p53. Progression into and through S phase was not altered by the presence of hypophosphorylated pRb in late G1, consistent with the failure of p53 to mediate G1 arrest in cells that are past the restriction point. These data indicate that accumulation of hypophosphorylated pRb has significantly different effects on cell cycle progression in early G1 versus late G1 or S phase.

DNA rereplication in the presence of mitotic spindle inhibitors in human and mouse fibroblasts lacking either p53 or pRb function

Cell cycle checkpoints are biochemical signal transduction pathways that prevent downstream events from being initiated until upstream processes are completed. We analyzed whether the p53 or pRb tumor suppressors are involved in a checkpoint(s) that prevents DNA rereplication in the presence of drugs that interfere with spindle assembly. Normal mouse and human fibroblasts arrested with a 4N DNA content when treated with nocodazole and Colcemid, whereas isogeneic p53-deficient or pRb-deficient derivatives became polyploid. Flow cytometric and cytogenetic analyses demonstrated that the polyploidy resulted from genome-wide rereplication without an intervening mitosis. Thus, p53 and pRb help maintain normal cell ploidy by preventing DNA rereplication prior to mitotic division.

p53 mediates permanent arrest over multiple cell cycles in response to gamma-irradiation

A new technique that monitors cell cycle progression over multiple cycles was used to gain insight into how p53 limits the emergence of variants with structural chromosome alterations following gamma-irradiation. G0-synchronized, p53+ (with a functional p53 pathway) normal human fibroblast and epithelial strains underwent a dose-dependent permanent arrest in the initial G0-G1 phase after irradiation. The dose-response curves indicate that a single event, such as an irreparable DNA break, may be sufficient to induce arrest. p53+ cells that escaped the initial G0-G1 phase after irradiation entered S phase in at least two waves. However, many of these cells underwent long-term arrest in subsequent phases. In contrast, virtually all of the cells in isogenic p53- (with a nonfunctional p53 pathway) strains escaped from the first G0-G1 phase without delay, regardless of the dose. p53- cells were also eliminated in subsequent phases but at significantly lower frequencies. Consistent with these findings, the reproductive viability of p53- cells was higher than p53+ cells. The nonclonogenic fraction appeared to be eliminated within three cycles for both cell types. In addition, artificial holding in G0 after irradiation, which allows for the repair of potentially lethal damage, led to similar increases in survival in p53+ and p53- cells. These data are inconsistent with the hypothesis that the primary function of p53-dependent G0-G1 arrest in response to gamma-irradiation is to allow additional time for DNA repair. Rather, they indicate that p53 helps maintain genetic stability by eliminating cells with damaged chromosomes from the reproductively viable population.

Positive selection of FLP-mediated unequal sister chromatid exchange products in mammalian cells

Site-specific recombination provides a powerful tool for studying gene function at predetermined chromosomal sites. Here we describe the use of a blasticidin resistance system to select for recombination in mammalian cells using the yeast enzyme FLP. The vector is designed so that site-specific recombination reconstructs the antibiotic resistance marker within the sequences flanked by the FLP target sites. This approach allows the detection of DNA excised by FLP-mediated recombination and facilitates the recovery of recombination products that would not be detected by available screening strategies. We used this system to show that the molecules excised by intrachromosomal recombination between tandem FLP recombinase target sites do not reintegrate into the host genome at detectable frequencies. We further applied the direct selection approach to recover a rare FLP-mediated recombination event displaying the characteristics of an unequal sister chromatid exchange between FLP target sites. Implications of this approach for the generation of duplications to assess their effect on gene dosage and chromosome stability are discussed.

The tumorigenic potential and cell growth characteristics of p53-deficient cells are equivalent in the presence or absence of Mdm2

The Mdm2 oncoprotein forms a complex with the p53 tumor suppressor protein and inhibits p53-mediated regulation of heterologous gene expression. Recently, Mdm2 has been found to bind several other proteins that function to regulate cell cycle progression, including the E2F-1/DP1 transcription factor complex and the retinoblastoma tumor-suppressor protein. To determine whether Mdm2 plays a role in cell cycle control or tumorigenesis that is distinct from its ability to modulate p53 function, we have examined and compared both the in vitro growth characteristics of p53-deficient and Mdm2/p53-deficient fibroblasts, and the rate and spectrum of tumor formation in p53-deficient and Mdm2/p53-deficient mice. We find no difference between p53-deficient fibroblasts and Mdm2/p53-deficient fibroblasts either in their rate of proliferation in culture or in their survival frequency when treated with various genotoxic agents. Cell cycle studies indicate no difference in the ability of the two cell populations to enter S phase when treated with DNA-damaging agents or nucleotide antimetabolites, and p53-deficient fibroblasts and Mdm2/p53-deficient fibroblasts exhibit the same rate of spontaneous immortalization following long-term passage in culture. Finally, p53-deficient mice and Mdm2/p53-deficient mice display the same incidence and spectrum of spontaneous tumor formation in vivo. These results demonstrate that deletion of Mdm2 has no additional effect on cell proliferation, cell cycle control, or tumorigenesis when p53 is absent.

Microsatellite instability correlates with reduced survival and poor disease prognosis in breast cancer

Size changes in microsatellite sequences have been detected in many types of cancer, but the influence of this form of genetic instability on disease progression remains unclear. We determined the incidence of microsatellite instability in breast cancer by comparing PCR-amplified sequences from paraffin-embedded samples of normal and tumor tissue from affected individuals. This analysis showed that at least 30% of breast cancers exhibit microsatellite instability (MI). Of importance, MI correlated with indicators commonly associated with poor disease prognosis, including lymph node status, tumor size, and advanced tumor stage. Individuals with MI+ tumors also showed significantly reduced disease-free and overall survival. These data contrast with studies showing that MI correlates with improved prognosis in colon and gastric cancers. We propose that defects resulting in MI promote disease progression and result in a poor prognosis in breast cancer.

p53-dependent cell cycle arrests are preserved in DNA-activated protein kinase-deficient mouse fibroblasts

p53 is involved in at least three cell cycle checkpoints in normal cells: two in G1, activated by either DNA damage or by ribonucleotide pool depletion in the absence of damage, and one in metaphase/anaphase activated by an incomplete mitotic spindle. We tested whether any of these checkpoints require the DNA-activated protein kinase (DNAPK), since data indicate that it is activated by damaged DNA to modify p53 in cultured cells and in cell-free systems. Fibroblasts isolated from mice with severe combined immune deficiency (SCID) were used because the sole genetic defect underlying this syndrome lies within the DNAPK gene. This report shows that age-matched SCID and isogenic wild-type embryonic fibroblasts arrested in response to DNA damage, ribonucleoside triphosphate depletion, and spindle poisons, whereas p53-/- fibroblasts failed to do so. Therefore, DNAPK-deficient scid cells preserve normal p53-dependent cell cycle checkpoints. The data provide one explanation of why scid mice are not tumor prone though they are deficient in double-strand break repair.

Sensitivity and selectivity of the DNA damage sensor responsible for activating p53-dependent G1 arrest

The tumor suppressor p53 contributes to maintaining genome stability by inducing a cell cycle arrest or apoptosis in response to conditions that generate DNA damage. Nuclear injection of linearized plasmid DNA, circular DNA with a large gap, or single-stranded circular phagemid is sufficient to induce a p53-dependent arrest. Supercoiled and nicked plasmid DNA, and circular DNA with a small gap were ineffective. Titration experiments indicate that the arrest mechanism in normal human fibroblasts can be activated by very few double strand breaks, and only one may be sufficient. Polymerase chain reaction assays showed that end-joining activity is low in serum-arrested human fibroblasts, and that higher joining activity occurs as cells proceed through G1 or into S phase. We propose that the exquisite sensitivity of the p53-dependent G1 arrest is partly due to inefficient repair of certain types of DNA damage in early G1.

A reversible, p53-dependent G0/G1 cell cycle arrest induced by ribonucleotide depletion in the absence of detectable DNA damage

Cells with a functional p53 pathway undergo a G0/G1 arrest or apoptosis when treated with gamma radiation or many chemotherapeutic drugs. It has been proposed that DNA damage is the exclusive signal that triggers the arrest response. However, we found that certain ribonucleotide biosynthesis inhibitors caused a p53-dependent G0 or early G1 arrest in the absence of replicative DNA synthesis or detectable DNA damage in normal human fibroblasts. CTP, GTP, or UTP depletion alone was sufficient to induce arrest. In contrast to the p53-dependent response to DNA damage, characterized by long-term arrest and irregular cellular morphologies, the antimetabolite-induced arrest was highly reversible and cellular morphologies remained relatively normal. Both arrest responses correlated with prolonged induction of p53 and the Cdk inhibitor P21(WAF1/CIP1/SDI1) and with dephosphorylation of pRb. Thus, we propose that p53 can serve as a metabolite sensor activated by depletion of ribonucleotides or products or processes dependent on ribonucleotides. Accordingly, p53 may play a role in inducing a quiescence-like arrest state in response to nutrient challenge and a senescence-like arrest state in response to DNA damage. These results have important implications for the mechanisms by which p53 prevents the emergence of genetic variants and for developing more effective approaches to chemotherapy based on genotype.

Selective capture of acentric fragments by micronuclei provides a rapid method for purifying extrachromosomally amplified DNA

The amplification and overexpression of a number of oncogenes is strongly associated with the progression of a variety of different cancers. We now present a strategy to purify amplified DNA on double minute chromosomes (DMs) to enable analysis of their prevalence and contribution to tumourigenesis. Using cell lines derived from four different tumour types, we have developed a general and rapid method to purify micronuclei that are known to entrap extrachromosomal elements. The isolated DNA is highly enriched in DM sequences and can be used to prepare probes to localize the progenitor single copy chromosomal regions. The capture of DMs by micronuclei appears to be dependent on their lack of a centromere rather than their small size.

Participation of the human beta-globin locus control region in initiation of DNA replication

The human beta-globin locus control region (LCR) controls the transcription, chromatin structure, and replication timing of the entire locus. DNA replication was found to initiate in a transcription-independent manner within a region located 50 kilobases downstream of the LCR in human, mouse, and chicken cells containing the entire human beta-globin locus. However, DNA replication did not initiate within a deletion mutant locus lacking the sequences that encompass the LCR. This mutant locus replicated in the 3' to 5' direction. Thus, interactions between distantly separated sequences can be required for replication initiation, and factors mediating this interaction appear to be conserved in evolution.

Lack of functional retinoblastoma protein mediates increased resistance to antimetabolites in human sarcoma cell lines

Growth inhibition assays indicated that the IC50 values for methotrexate (MTX) and 5-fluorodeoxyuridine (FdUrd) in HS-18, a liposarcoma cell line lacking retinoblastoma protein (pRB), and SaOS-2, an osteosarcoma cell line with a truncated and nonfunctional pRB, were 10- to 12-fold and 4- to 11-fold higher, respectively, than for the HT-1080 (fibrosarcoma) cell line, which has wild-type pRB. These Rb-/- cell lines exhibited a 2- to 4-fold increase in both dihydrofolate reductase (DHFR) and thymidylate synthase (TS) enzyme activities as well as a 3- to 4-fold increase in mRNA levels for these enzymes compared to the HT-1080 (Rb+/+) cells. This increase in expression was not due to amplification of the DHFR and TS genes. Growth inhibition by MTX and FdUrd was increased and DHFR and TS activities and expression were correspondingly decreased in Rb transfectants of SaOS-2 cells. In contrast, there was no significant difference in growth inhibition among these cell lines for the nonantimetabolites VP-16, cisplatin, and doxorubicin. A gel mobility-shift assay showed that parental SaOS-2 cells had increased levels of free E2F compared to the Rb-reconstituted SaOS-2 cells. These results indicate that pRB defective cells may have decreased sensitivity to growth inhibition by target enzymes encoded by genes whose transcription is enhanced by E2F proteins and suggest mechanisms of interaction between cytotoxic agents and genes involved in cell cycle progression.

Identification of an origin of bidirectional DNA replication in the ubiquitously expressed mammalian CAD gene

Most DNA replication origins in eukaryotes localize to nontranscribed regions, and there are no reports of origins within constitutively expressed genes. This observation has led to the proposal that there may be an incompatibility between origin function and location within a ubiquitously expressed gene. The biochemical and functional evidence presented here demonstrates that an origin of bidirectional replication (OBR) resides within the constitutively expressed housekeeping gene CAD, which encodes the first three reactions of de novo uridine biosynthesis (carbamoyl-phosphate synthetase, aspartate carbamoyltransferase, and dihydroorotase). The OBR was localized to a 5-kb region near the center of the Syrian hamster CAD transcriptional unit. DNA replication initiates within this region in the single-copy CAD gene in Syrian baby hamster kidney cells and in the large chromosomal amplicons that were generated after selection with N-phosphonacetyl-L-aspartate, a specific inhibitor of CAD. DNA synthesis also initiates within this OBR in autonomously replicating extrachromosomal amplicons (CAD episomes) located in an N-phosphonacetyl-L-aspartate-resistant clone (5P20) of CHOK1 cells. CAD episomes consist entirely of a multimer of Syrian hamster CAD cosmid sequences (cCAD1). These data limit the functional unit of replication initiation and timing control to the 42 kb of Syrian hamster sequences contained in cCAD1. In addition, the data indicate that the origin recognition machinery is conserved across species, since the same OBR region functions in both Syrian and Chinese hamster cells. Importantly, while cCAD1 exhibits characteristics of a complete replicon, we have not detected autonomous replication directly following transfection. Since the CAD episome was generated after excision of chromosomally integrated transfected cCAD1 sequences, we propose that prior localization within a chromosome may be necessary to "license" some biochemically defined OBRs to render them functional.

Deficiency of retinoblastoma protein leads to inappropriate S-phase entry, activation of E2F-responsive genes, and apoptosis

The retinoblastoma susceptibility gene (Rb) participates in controlling the G1/S-phase transition, presumably by binding and inactivating E2F transcription activator family members. Mouse embryonic fibroblasts (MEFs) with no, one, or two inactivated Rb genes were used to determine the specific contributions of Rb protein to cell cycle progression and gene expression. MEFs lacking both Rb alleles (Rb-/-) entered S phase in the presence of the dihydrofolate reductase inhibitor methotrexate. Two E2F target genes, dihydrofolate reductase and thymidylate synthase, displayed elevated mRNA and protein levels in Rb- MEFs. Since absence of functional Rb protein in MEFs is sufficient for S-phase entry under growth-limiting conditions, these data indicate that the E2F complexes containing Rb protein, and not the Rb-related proteins p107 and p130, may be rate limiting for the G1/S transition. Antineoplastic drugs caused accumulation of p53 in the nuclei of both Rb+/+ and Rb-/- MEFs. While p53 induction led to apoptosis in Rb-/- MEFs, Rb+/- and Rb+/+ MEFs underwent cell cycle arrest without apoptosis. These results reveal that diverse growth signals work through Rb to regulate entry into S phase, and they indicate that absence of Rb protein produces a constitutive DNA replication signal capable of activating a p53-associated apoptotic response.

Genetic instability as a consequence of inappropriate entry into and progression through S-phase

The stability of the mammalian genome depends on the proper function of G1 and G2 cell cycle control mechanisms. Two tumor suppressors, p53 and retinoblastoma (Rb), play key roles in progression from G1 into S-phase. We address the mechanisms by which these proteins mediate a G1 arrest in response to DNA damage and limiting metabolic conditions. Gamma-irradiation induced a prolonged, p53-dependent G1 arrest associated with a long-term increase in the levels of the cdk-inhibitor p21WAFl/Cipl (p21). Microinjection of linear plasmid DNA also caused a G1 arrest. The p53-dependent arrest induced by inhibitors of UMP biosynthesis was reversible and occurred in the absence of detectable DNA damage. Both arrest mechanisms contribute to limiting the formation and propagation of damaged genomes. Cells containing mutant p53 but wild-type Rb do not generate methotrexate (Mtx) resistant variants. However, pre-treatment with DNA damaging agents prior to drug selection resulted in resistant clones containing amplified dihydrofolate reductase (DHFR) genes, suggesting that DNA breakage is a rate limiting step for gene amplification. The Mtx-induced arrest did not occur in cells with non-functional Rb. Rb acts as a negative regulator of the E2F transcription factors, and Rb-deficient primary mouse embryo fibroblasts (MEFs) produced elevated levels of mRNA and protein for key E2F target genes. Failure to prevent entry into S-phase in Rb-/- MEFs exposed to DNA-damaging or nutrient limiting conditions caused apoptosis and correlated with p53 induction. Taken together, these findings indicate a link between p53 and Rb function and suggest that their coordination insures correct entry into S-phase, minimizing the emergence of genetic variants.

DNA damage triggers a prolonged p53-dependent G1 arrest and long-term induction of Cip1 in normal human fibroblasts

The tumor suppressor p53 is a cell cycle checkpoint protein that contributes to the preservation of genetic stability by mediating either a G1 arrest or apoptosis in response to DNA damage. Recent reports suggest that p53 causes growth arrest through transcriptional activation of the cyclin-dependent kinase (Cdk)-inhibitor Cip1. Here, we characterize the p53-dependent G1 arrest in several normal human diploid fibroblast (NDF) strains and p53-deficient cell lines treated with 0.1-6 Gy gamma radiation. DNA damage and cell cycle progression analyses showed that NDF entered a prolonged arrest state resembling senescence, even at low doses of radiation. This contrasts with the view that p53 ensures genetic stability by inducing a transient arrest to enable repair of DNA damage, as reported for some myeloid leukemia lines. Gamma radiation administered in early to mid-, but not late, G1 induced the arrest, suggesting that the p53 checkpoint is only active in G1 until cells commit to enter S phase at the G1 restriction point. A log-linear plot of the fraction of irradiated G0 cells able to enter S phase as a function of dose is consistent with single-hit kinetics. Cytogenetic analyses combined with radiation dosage data indicate that only one or a small number of unrepaired DNA breaks may be sufficient to cause arrest. The arrest also correlated with long-term elevations of p53 protein, Cip1 mRNA, and Cip1 protein. We propose that p53 helps maintain genetic stability in NDF by mediating a permanent cell cycle arrest through long-term induction of Cip1 when low amounts of unrepaired DNA damage are present in G1 before the restriction point.

Pharmacokinetics of hydroxyurea in nude mice

Extrachromosomal DNA is the predominant form of gene amplification in human tumors. Hydroxyurea (HU) concentrations of 100-150 microM have been promising in vitro for extrachromosomal DNA elimination. The study objective was to determine the HU dose-concentration relationship in nude mice with HU doses from 0 to 200 mg/kg. For HU t1/2 determination, mice were injected with HU 100 mg/kg. A plasma concentration of 159 microM was achieved and a t1/2 of 11.3 min determined. Based on these findings, In vivo elimination studies will require frequent administration of HU to maintain plasma concentrations from 100 to 150 microM.

Induction of differentiation in HL60 cells by the reduction of extrachromosomally amplified c-myc

Oncogene amplification in tumor cells results in the overexpression of proteins that confer a growth advantage in vitro and in vivo. Amplified oncogenes can reside intrachromosomally, within homogeneously staining regions (HSRs), or extrachromosomally, within double minute chromosomes (DMs). Since previous studies have shown that low concentrations of hydroxyurea (HU) can eliminate DMs, we studied the use of HU as a gene-targeting agent in tumor cells containing extrachromosomally amplified oncogenes. In a neuroendocrine cell line (COLO 320), we have shown that HU can eliminate amplified copies of c-myc located on DMs, leading to a reduction in tumorigenicity in vitro and in vivo. To determine whether the observed reduction in tumorigenicity was due to differentiation, we next investigated whether HU could induce differentiation in HL60 cells containing extrachromosomally amplified c-myc. We compared the effects of HU, as well as two other known differentiating agents (dimethyl sulfoxide and retinoic acid), on c-myc gene copy number, c-myc expression, and differentiation in HL60 cells containing amplified c-myc genes either on DMs or HSRs. We discovered that HU and dimethyl sulfoxide reduced both c-myc gene copy number and expression and induced differentiation in cells containing c-myc amplified on DMs. These agents failed to have similar effects on HL60 cells with amplified c-myc in HSRs. By contrast, retinoic acid induced differentiation independent of the localization of amplified c-myc. These data illustrate the utility of targeting extrachromosomal DNA to modulate tumor phenotype and reveal that both HU and dimethyl sulfoxide induce differentiation in HL60 cells through DM elimination.

Chromosomal localization of the human retinoid X receptors

The recently described retinoid X receptors (RXRs) respond to the novel retinoid 9-cis-retinoic acid and also serve as heterodimeric partners for the vitamin D, thyroid hormone, and retinoic acid receptors (VDR, TR, and RAR, respectively). In this work, we report high-resolution localization of the human RXR genes within cytogenetic bands and also within a standard reference map of cosmid DNA markers on human chromosomes. We have determined the location of the human RXR genes by pairwise hybridization of the RXR cosmids and reference markers, using fluorescence in situ hybridization. We localized (i) RXR alpha (RXRA) to chromosome 9 band q34.3; (ii) RXR beta (RXRB) to chromosome 6 band 21.3; and (iii) RXR gamma (RXRG) to chromosome 1 band q22-q23. Six retinoid-responsive transcription factors have been identified so far, including three retinoic acid receptors in addition to the three RXRs. Interestingly, each of these receptors in human and mouse is encoded by genes located at distinct chromosomal loci and on separate chromosomes. The proximity of RXR genes to loci known to be associated with genetic disorders suggests that their location may be useful in establishing a link between RXRs and certain human diseases.

Localization of a bidirectional DNA replication origin in the native locus and in episomally amplified murine adenosine deaminase loci

Gene amplification is frequently mediated by the initial production of acentric, autonomously replicating extrachromosomal elements. The 4,000 extrachromosomal copies of the mouse adenosine deaminase (ADA) amplicon in B-1/50 cells initiate their replication remarkably synchronously in early S phase and at approximately the same time as the single-copy chromosomal locus from which they were derived. The abundance of ADA sequences and favorable replication timing characteristics in this system led us to determine whether DNA replication initiates in ADA episomes within a preferred region and whether this region is the same as that used at the corresponding chromosomal locus prior to amplification. This study reports the detection and localization of a discrete set of DNA fragments in the ADA amplicon which label soon after release of synchronized B-1/50 cells into S phase. A switch in template strand complementarity of Okazaki fragments, indicative of the initiation of bidirectional DNA replication, was found to lie within the same region. This putative replication origin is located approximately 28.5 kbp upstream of the 5' end of the ADA gene. The same region initiated DNA replication in the single-copy ADA locus of the parental cells. These analyses provide the first evidence that the replication of episomal intermediates involved in gene amplification initiates within a preferred region and that the same region is used to initiate DNA synthesis within the native locus.

Molecular dissection of an extrachromosomal amplicon reveals a circular structure consisting of an imperfect inverted duplication

A mouse fibroblast line, B-1/50, with a 4300-fold amplification of the adenosine deaminase gene locus (Yeung et al., 1983, J. Biol. Chem. 258: 8338-8345), was shown by in situ hybridization to harbor the amplified sequences on variously sized extrachromosomal elements. We show here that the smallest circle is approximately 500 kb. We describe a facile screening technique for identifying cosmid and yeast artificial chromosome (YAC) clones derived from the amplicon. A closed molecular map was generated by arranging the cosmids and YACs into a contig spanning over 250 kb of the adenosine deaminase gene locus. YACs from the two ends of this contig were shown to delimit a 250-kb inverted duplication. Long-range mapping of a SalI partial digest of B-1/50 DNA is also consistent with the interpretation that the 500-kb adenosine deaminase amplicon in B-1/50 cells is an inverted duplication. The finding that this amplicon is the only or predominant structure containing amplified sequences in the B-1/50 cell line suggests that such structures are not inherently prone to high frequency rearrangement, even when present at such high copy number. This study provides the first molecular description of the structure of an episome involved in mammalian gene amplification. The implications of this finding for models of gene amplification and episome formation are discussed.