Adenovirus type 12-induced fragility of the human RNU2 locus requires U2 small nuclear RNA transcriptional regulatory elements

Topological epigenetics: The biophysics of DNA supercoiling and its relation to transcription and genome instability

Whilst DNA encodes our genetic blueprint as individual nucleobases, as well as epigenetic annotations in the form of biochemical marks, it also carries an extra layer of topological information -, the local over or underwinding of the double helix, known as DNA supercoiling. Supercoiling is a fundamental property of DNA that can be viewed as "topological epigenetics": it stores energy and structural information, and is tightly linked to fundamental processes; however, its quantification and study, by experiments and modelling alike, is challenging. We review experimental and simulation techniques to study supercoiling and its partition into twist and writhe, especially in the context of chromatin. We then discuss the dynamics of transcription-driven supercoiling in vitro and in vivo, and of supercoiling propagation along mammalian genomes. We finally provide evidence from the literature and potential mechanisms linking this ethereal topological mark to gene expression and chromosome instabilities in genetic diseases and cancer.

The Promotion of Genomic Instability in Human Fibroblasts by Adenovirus 12 Early Region 1B 55K Protein in the Absence of Viral Infection

The adenovirus 12 early region 1B55K (Ad12E1B55K) protein has long been known to cause non-random damage to chromosomes 1 and 17 in human cells. These sites, referred to as Ad12 modification sites, have marked similarities to classic fragile sites. In the present report we have investigated the effects of Ad12E1B55K on the cellular DNA damage response and on DNA replication, considering our increased understanding of the pathways involved. We have compared human skin fibroblasts expressing Ad12E1B55K (55K+HSF), but no other viral proteins, with the parental cells. Appreciable chromosomal damage was observed in 55K+HSFs compared to parental cells. Similarly, an increased number of micronuclei was observed in 55K+HSFs, both in cycling cells and after DNA damage. We compared DNA replication in the two cell populations; 55K+HSFs showed increased fork stalling and a decrease in fork speed. When replication stress was introduced with hydroxyurea the percentage of stalled forks and replication speeds were broadly similar, but efficiency of fork restart was significantly reduced in 55K+HSFs. After DNA damage, appreciably more foci were formed in 55K+HSFs up to 48 h post treatment. In addition, phosphorylation of ATM substrates was greater in Ad12E1B55K-expressing cells following DNA damage. Following DNA damage, 55K+HSFs showed an inability to arrest in cell cycle, probably due to the association of Ad12E1B55K with p53. To confirm that Ad12E1B55K was targeting components of the double-strand break repair pathways, co-immunoprecipitation experiments were performed which showed an association of the viral protein with ATM, MRE11, NBS1, DNA-PK, BLM, TOPBP1 and p53, as well as with components of the replisome, MCM3, MCM7, ORC1, DNA polymerase δ, TICRR and cdc45, which may account for some of the observed effects on DNA replication. We conclude that Ad12E1B55K impacts the cellular DNA damage response pathways and the replisome at multiple points through protein-protein interactions, causing genomic instability.

Direct visualization of the highly polymorphic RNU2 locus in proximity to the BRCA1 gene

Although the breast cancer susceptibility gene BRCA1 is one of the most extensively characterized genetic loci, much less is known about its upstream variable number tandem repeat element, the RNU2 locus. RNU2 encodes the U2 small nuclear RNA, an essential splicing element, but this locus is missing from the human genome assembly due to the inherent difficulty in the assembly of repetitive sequences. To fill the gap between RNU2 and BRCA1, we have reconstructed the physical map of this region by re-examining genomic clone sequences of public databases, which allowed us to precisely localize the RNU2 array 124 kb telomeric to BRCA1. We measured by performing FISH analyses on combed DNA for the first time the exact number of repeats carried by each of the two alleles in 41 individuals and found a range of 6-82 copies and a level of heterozygosity of 98%. The precise localisation of the RNU2 locus in the genome reference assembly and the implementation of a new technical tool to study it will make the detailed exploration of this locus possible. This recently neglected macrosatellite could be valuable for evaluating the potential role of structural variations in disease due to its location next to a major cancer susceptibility gene.

Transcriptional regulation of human small nuclear RNA genes

The products of human snRNA genes have been frequently described as performing housekeeping functions and their synthesis refractory to regulation. However, recent studies have emphasized that snRNA and other related non-coding RNA molecules control multiple facets of the central dogma, and their regulated expression is critical to cellular homeostasis during normal growth and in response to stress. Human snRNA genes contain compact and yet powerful promoters that are recognized by increasingly well-characterized transcription factors, thus providing a premier model system to study gene regulation. This review summarizes many recent advances deciphering the mechanism by which the transcription of human snRNA and related genes are regulated.

Human U2 snRNA genes exhibit a persistently open transcriptional state and promoter disassembly at metaphase

In mammals, small multigene families generate spliceosomal U snRNAs that are nearly as abundant as rRNA. Using the tandemly repeated human U2 genes as a model, we show by footprinting with DNase I and permanganate that nearly all sequences between the enhancer-like distal sequence element and the initiation site are protected during interphase whereas the upstream half of the U2 snRNA coding region is exposed. We also show by chromatin immunoprecipitation that the SNAPc complex, which binds the TATA-like proximal sequence element, is removed at metaphase but remains bound under conditions that induce locus-specific metaphase fragility of the U2 genes, such as loss of CSB, BRCA1, or BRCA2 function, treatment with actinomycin D, or overexpression of the tetrameric p53 C terminus. We propose that the U2 snRNA promoter establishes a persistently open state to facilitate rapid reinitiation and perhaps also to bypass TFIIH-dependent promoter melting; this open state would then be disassembled to allow metaphase chromatin condensation.

The p53 tumor suppressor protein represses human snRNA gene transcription by RNA polymerases II and III independently of sequence-specific DNA binding

Human U1 and U6 snRNA genes are transcribed by RNA polymerases II and III, respectively. While the p53 tumor suppressor protein is a general repressor of RNA polymerase III transcription, whether p53 regulates snRNA gene transcription by RNA polymerase II is uncertain. The data presented herein indicate that p53 is an effective repressor of snRNA gene transcription by both polymerases. Both U1 and U6 transcription in vitro is repressed by recombinant p53, and endogenous p53 occupancy at these promoters is stimulated by UV light. In response to UV light, U1 and U6 transcription is strongly repressed. Human U1 genes, but not U6 genes, contain a high-affinity p53 response element located within the core promoter region. Nonetheless, this element is not required for p53 repression and mutant p53 molecules that do not bind DNA can maintain repression, suggesting a reliance on protein interactions for p53 promoter recruitment. Recruitment may be mediated by the general transcription factors TATA-box binding protein and snRNA-activating protein complex, which interact well with p53 and function for both RNA polymerase II and III transcription.

Viral induction of site-specific chromosome damage

The advent of advanced cell culture and cytogenetics techniques in the 1950s opened a new avenue for research on the pathogenic interactions between animal viruses and their hosts. Studies of many viruses revealed their ability to nonspecifically induce cytogenetic damage to their host cell's chromosomes. However, only three viruses, the oncogenic adenoviruses, herpes simplex virus (HSV) and human cytomegalovirus (HCMV), have been found to cause non-random, site-specific chromosomal damage. Adenovirus (Ad) type 12 induces fragility at four distinct loci (RNU1, RNU2, RN5S and PSU1) in many different types of human cells. A common feature of these loci is that they contain a repeated array of transcriptionally active genes encoding small structural RNAs. Site-specific induction of breaks also requires the virally encoded E1B protein of M(r) 55000 and the C-terminus of the cellular p53 protein. Analysis of the induction of damage by HSV and HCMV necessitates consideration of several factors, including the strain of virus used, the timing of infection, the type of cell used, and the multiplicity of infection. Both HSV strains 1 and 2 are cytotoxic, although the former seems to be more proficient at inducing damage. At early times post infection, HSV induces breaks and specific uncoiling of the centromeres of chromosomes 1, 9 and 16. This is followed at later times by a more complete severing of all of the chromosomes, termed pulverisation. Damage by HSV requires viral entry and de novo viral protein synthesis, with immediate early viral proteins responsible for the induction of breaks and uncoiling and early gene products (most likely nucleases) involved in the extensive pulverisation seen later. HCMV has been studied primarily in permissive human fibroblasts. Its ability to induce specific damage in chromosome 1 at two loci, 1q21 and 1q42, was only recently revealed as the cells must be in S-phase when they are infected for the breaks to be observed. In contrast to adenovirus and HSV, HCMV induction of specific breakage requires only viral entry into the cell and not de novo viral protein expression. This latter point may be a factor in its ability to cause damage in the developing fetal brain, where the most severe clinical manifestations of congenital infection are observed.

RNA-mediated interaction of Cajal bodies and U2 snRNA genes

Cajal bodies (CBs) are nuclear structures involved in RNA metabolism that accumulate high concentrations of small nuclear ribonucleoproteins (snRNPs). Notably, CBs preferentially associate with specific genomic loci in interphase human cells, including several snRNA and histone gene clusters. To uncover functional elements involved in the interaction of genes and CBs, we analyzed the expression and subcellular localization of stably transfected artificial arrays of U2 snRNA genes. Although promoter substitution arrays colocalized with CBs, constructs containing intragenic deletions did not. Additional experiments identified factors within CBs that are important for association with the native U2 genes. Inhibition of nuclear export or targeted degradation of U2 snRNPs caused a marked decrease in the levels of U2 snRNA in CBs and strongly disrupted the interaction with U2 genes. Together, the results illustrate a specific requirement for both the snRNA transcripts as well as the presence of snRNPs (or snRNP proteins) within CBs. Our data thus provide significant insight into the mechanism of CB interaction with snRNA loci, strengthening the putative role for this nuclear suborganelle in snRNP biogenesis.

d(GA x TC)(n) microsatellite DNA sequences enhance homologous DNA recombination in SV40 minichromosomes

The genomic distribution of the abundant eukaryotic d(GA x TC)(n) DNA microsatellite suggests that it could contribute to DNA recombination. Here, it is shown that this type of microsatellite DNA sequence enhances DNA recombination in SV40 minichromosomes, the rate of homologous DNA recombination increasing by as much as two orders of magnitude in the presence of a d(GA x TC)(22) sequence. This effect depends on the region of the SV40 genome at which the d(GA x TC)(22) sequence is cloned. It is high when the sequence is located proximal to the SV40 control region but no effect is observed when located 3.5 kb away from the SV40 ori. These results indicate that the recombination potential of d(GA x TC)(n) sequences is likely linked to DNA replication and/or transcription. The potential contribution of the structural properties of d(GA x TC)(n) sequences to this effect is discussed.

Activation of p53 or loss of the Cockayne syndrome group B repair protein causes metaphase fragility of human U1, U2, and 5S genes

Infection by adenovirus 12, transfection with the Ad12 E1B 55 kDa gene, or activation of p53 cause metaphase fragility of four loci (RNU1, PSU1, RNU2, and RN5S) each containing tandemly repeated genes for an abundant small RNA (U1, U2, and 5S RNA). We now show that loss of the Cockayne syndrome group B protein (CSB) or overexpression of the p53 carboxy-terminal domain induces fragility of the same loci; moreover, p53 interacts with CSB in vivo and in vitro. We propose that CSB functions as an elongation factor for transcription of structured RNAs, including some mRNAs. Activation of p53 would inhibit CSB, stalling transcription complexes and locally blocking chromatin condensation. Impaired transcription elongation may also explain the diverse clinical features of Cockayne syndrome.

E1B 55-kilodalton oncoproteins of adenovirus types 5 and 12 inactivate and relocalize p53, but not p51 or p73, and cooperate with E4orf6 proteins to destabilize p53

The p53 tumor suppressor protein represents a target for viral and cellular oncoproteins, including adenovirus gene products. Recently, it was discovered that several proteins with structural and functional homologies to p53 exist in human cells. Two of them were termed p51 and p73. We have shown previously that the E1B 55-kDa protein (E1B-55 kDa) of adenovirus type 5 (Ad5) binds and inactivates p53 but not p73. Further, p53 is rapidly degraded in the presence of E1B-55 kDa and the E4orf6 protein of this virus. Here, it is demonstrated that p51 does not detectably associate with E1B-55 kDa. While p53 is relocalized to the cytoplasm by E1B-55 kDa, p51's location is unaffected. Finally, p51 retains its full transcriptional activity in the presence of E1B-55 kDa. Apparently, p51 does not represent a target of Ad5 E1B-55 kDa, suggesting that the functions of p51 are distinct from p53-like tumor suppression. E1B-55 kDa from highly oncogenic adenovirus type 12 (Ad12) was previously shown to surpass the oncogenic activity of Ad5 E1B-55 kDa in various assay systems, raising the possibility that Ad12 E1B-55 kDa might target a broader range of p53-like proteins. However, we show here that Ad12 E1B-55 kDa also inhibits p53's transcriptional activity without measurably affecting p73 or p51. Moderate inhibition of p51's transcriptional activity was observed in the presence of the E4orf6 proteins from Ad5 and Ad12. p53 and Ad12-E1B-55 kDa colocalize in the nucleus and also in cytoplasmic clusters when transiently coexpressed. Finally, E1B-55 kDa and E4orf6 of Ad12 mediate rapid degradation of p53 with an efficiency comparable to that of the Ad5 proteins in human and rodent cells. Our results suggest that E1B-55 kDa of either virus type has similar effects on p53 but does not affect p73 and p51.

Striking bimodal methylation of the repeat unit of the tandem array encoding human U2 snRNA (the RNU2 locus)

The genes encoding human U2 small nuclear RNA are arrayed in tandem (the RNU2 locus) and have undergone concerted evolution for >35 Myr. Tandem organization of repetitive sequences may facilitate recombination that underlies concerted evolution, but could risk instability. Since DNA methylation plays a crucial role in genome stability, we investigated the methylation status of the RNU2 locus to understand the forces maintaining array stability and homogeneity. We found that a region of approximately 1.5 kb spanning the U2 promoter, U2 gene sequence, and CT microsatellite is completely unmethylated, whereas the rest of the repeat is heavily methylated. Since the U2 transcription enhancer DSE and CT microsatellite mark the boundaries between methylated and unmethylated domains, they might function as cis-acting elements for establishing and maintaining proper methylation at the RNU2 locus. Interestingly, the RNU2 locus in human fibrosarcoma line HT1080 is hypomethylated, and de novo methylation did not occur in an artificial U2 tandem array introduced by stable transfection. The observed bimodal methylation pattern may be important for both efficient transcription of U2 gene and maintenance of nearly perfect tandem arrays in somatic cells.

Association of snRNA genes with coiled bodies is mediated by nascent snRNA transcripts

BACKGROUND

Coiled bodies are nuclear organelles that are highly enriched in small nuclear ribonucleoproteins (snRNPs) and certain basal transcription factors. Surprisingly, coiled bodies not only contain mature U snRNPs but also associate with specific chromosomal loci, including gene clusters that encode U snRNAs and histone messenger RNAs. The mechanism(s) by which coiled bodies associate with these genes is completely unknown.

RESULTS

Using stable cell lines, we show that artificial tandem arrays of human U1 and U2 snRNA genes colocalize with coiled bodies and that the frequency of the colocalization depends directly on the transcriptional activity of the array. Association of the genes with coiled bodies was abolished when the artificial U2 arrays contained promoter mutations that prevent transcription or when RNA polymerase II transcription was globally inhibited by alpha-amanitin. Remarkably, the association was also abolished when the U2 snRNA coding regions were replaced by heterologous sequences.

CONCLUSIONS

The requirement for the U2 snRNA coding region indicates that association of snRNA genes with coiled bodies is mediated by the nascent U2 RNA itself, not by DNA or DNA-bound proteins. Our data provide the first evidence that association of genes with a nuclear organelle can be directed by an RNA and suggest an autogenous feedback regulation model.

Direct cloning and analysis of DNA sequences from a region of the Chinese hamster genome associated with aphidicolin-sensitive fragility

Fragile sites are reproducibly expressed and chemically induced decondensations on mitotic chromosomes observed under cytological conditions. They are classified both on the basis of the frequency with which they occur (rare and common) and in terms of the chemical agent used to induce expression in tissue culture cells. Aphidicolin-sensitive common fragile sites appear to be ubiquitous in humans and other mammals and have been considered as candidates of pathological importance. Recently DNA from FRA3B, the most highly expressed constitutive fragile site in the human genome, has been cloned although as yet the cause of the underlying fragility has not been identified. In this study we describe the isolation, using a direct cloning approach, of DNA from a region of the Chinese hamster genome associated with aphidicolin-inducible fragility. Cells of a human-hamster somatic cell hybrid were transfected with a pSV2HPRT vector while exposed to aphidicolin, an inhibitor of DNA polymerases alpha, delta and epsilon. FISH analysis of stable transfectant clones revealed that the ingoing plasmid DNA had preferentially integrated into fragile site-containing chromosomal bands. Plasmid rescue was used to recover DNA sequences flanking one such integration site in the hamster genome. We demonstrate by FISH analysis of metaphase cells induced with aphidicolin that the rescued DNA is from a region of fragility on Chinese hamster chromosome 2, distal to the DHFR locus. Analysis of the DNA sequences flanking the integration site revealed the overall A+T content of the 3,725 bp region sequenced to be 63.3%, with a highly [A].[T]-rich 156 bp region (86.5%) almost adjacent to the integration site. Computational analyses have identified strong homologies to Saccharomyces cerevisiae autonomous replicating sequences (ARS), polypyrimidine tracts, scaffold attachment site consensus sequences and a 24 bp consensus sequence highly conserved in eukaryotic replication origins, all of which appear to cluster around the [A].[T]-rich sequences. This domain also possesses structural characteristics which are common to both prokaryotic and eukaryotic origins of replications, in particular an unusually straight conformation of low thermal stability flanked either side by highly bent DNA segments. Further isolation and characterisation of DNA sequences from common fragile sites will facilitate studies into the underlying nature of these enigmatic regions of the mammalian genome, leading to a greater understanding of chromatin structure.

A tandem array of minimal U1 small nuclear RNA genes is sufficient to generate a new adenovirus type 12-inducible chromosome fragile site

Infection of human cells with adenovirus serotype 12 (Ad12) induces metaphase fragility of four, and apparently only four, chromosomal loci. Surprisingly, each of these four loci corresponds to a cluster of genes encoding a small abundant structural RNA: the RNU1 and RNU2 loci contain tandemly repeated genes encoding U1 and U2 small nuclear RNAs (snRNAs), respectively; the PSU1 locus is a cluster of degenerate U1 genes; and the RN5S locus contains the tandemly repeated genes encoding 5S rRNA. These observations suggested that high local levels of transcription, in combination with Ad12 early functions, can interfere with metaphase chromatin packing. In support of this hypothesis, we and others found that an artificial tandem array of transcriptionally active, but not inactive, U2 snRNA genes would generate a novel Ad12-inducible fragile site. Although U1 and U2 snRNA are both transcribed by RNA polymerase II and share similar enhancer, promoter, and terminator signals, the human U1 promoter is clearly more complex than that of U2. In addition, the natural U1 tandem repeat unit exceeds 45 kb, whereas the U2 tandem repeat unit is only 6.1 kb. We therefore asked whether an artificial array of minimal U1 genes would also generate a novel Ad12-inducible fragile site. The exogenous U1 genes were marked by an innocuous U72C point mutation within the U1 coding region so that steady-state levels of U1 snRNA derived from the artificial array could be quantified by a simple primer extension assay. We found that the minimal U1 genes were efficiently expressed and were as effective as minimal U2 genes in generating a novel Ad12-inducible fragile site. Thus, despite significant differences in promoter architecture and overall gene organization, the active U1 transcription units suffice to generate a new virally inducible fragile site.

Adenovirus type 12-induced fragility of the human RNU2 locus requires p53 function

Adenovirus type 12 (Ad12) infection of human cells induces four chromosomal fragile sites corresponding to the U1 small nuclear RNA (snRNA) genes (the RNU1 locus), the U2 snRNA genes (RNU2), the U1 snRNA pseudogenes (PSU1), and the 5S rRNA genes (RN5S). Ad12-induced fragility of the RNU2 locus requires U2 snRNA transcriptional regulatory elements and viral early functions but not viral replication or integration, or chromosomal sequences flanking the RNU2 locus. We now show that Ad12 cannot induce the RNU1, RNU2, or PSU1 fragile sites in Saos-2 cells lacking the p53 and retinoblastoma (Rb) proteins but that viral induction of fragility is rescued in these cells when the expression of wild-type p53 or selected hot-spot mutants (i.e., V143A, R175H, R248W, and R273H) is restored by transient expression or stable retroviral transduction. We also observed weak constitutive fragility of the RNU1 and RNU2 loci in cells belonging to xeroderma pigmentosum complementation groups B and D (XPB and XPD) which are partially defective in the ERCC2 (XPD) and ERCC3 (XPB) helicase activities shared between the repairosome and the RNA polymerase H basal transcription factor TFIIH. We propose a model for Ad12-induced chromosome fragility in which interaction of p53 with the Ad12 E1B 55-kDa transforming protein (and possibly E4orf6) induces a p53 gain of function which ultimately perturbs the RNA polymerase II basal transcription apparatus. The p53 gain of function could interfere with chromatin condensation either by blocking mitotic shutdown of U1 and U2 snRNA transcription or by phenocopying global or local DNA damage. Specific fragilization of the RNU1, RNU2, and PSU1 loci could reflect the unusually high local concentration of strong transcription units or the specialized nature of the U1 and U2 snRNA transcription apparatus.

The microsatellite sequence (CT)n x (GA)n promotes stable chromosomal integration of large tandem arrays of functional human U2 small nuclear RNA genes

The multigene family encoding human U2 small nuclear RNA (snRNA) is organized as a single large tandem array containing 5 to 25 copies of a 6.1-kb repeat unit (the RNU2 locus). Remarkably, each of the repeat units within an individual U2 tandem array appears to be identical except for an irregular dinucleotide tract, known as the CT microsatellite, which exhibits minor length and sequence polymorphism. Using a somatic cell genetic assay, we previously noticed that the CT microsatellite appeared to stabilize artificial tandem arrays of U2 snRNA genes. We now demonstrate that the CT microsatellite is required to establish large tandem arrays of transcriptionally active U2 genes, increasing both the average and maximum size of the resulting arrays. In contrast, the CT microsatellite has no effect on the average or maximal size of artificial arrays containing transcriptionally inactive U2 genes that lack key promoter elements. Our data reinforce the connection between recombination and transcription. Active U2 transcription interferes with establishment or maintenance of the U2 tandem array, and the CT microsatellite opposes these effects, perhaps by binding GAGA or GAGA-related factors which alter local chromatin structure. We speculate that the mechanisms responsible for maintenance of tandem arrays containing active promoters may differ from those that maintain tandem arrays of transcriptionally inactive sequences.

Induction of fragility at the human RNU2 locus by cytosine arabinoside is dependent upon a transcriptionally competent U2 small nuclear RNA gene and the expression of p53

Chromosomal fragile sites are regions that are intrinsically unstable and are susceptible to experimentally induced damage. In most cases, the target and mechanism of induction of fragility are unknown. Using ectopic integration of engineered DNA arrays to create "new" fragile sites, we and others have previously shown that the transcriptionally competent U2 gene is necessary and sufficient for induction of fragility at the RNU2 locus upon infection of human cells with Adenovirus 12. In the present study we have investigated the response of the RNU2 locus to cytosine arabinoside (araC), an inhibitor of DNA polymerases and a common inducer of fragile sites. We demonstrate that the RNU2 locus is sensitive to the drug and that araC-induced fragility is dependent upon a functional U2 gene and on the expression of the cellular p53 protein. Our results identify a novel DNA structure associated with fragile sites and suggest a role for transcription and repair processes in RNU2 fragility.

Genes transfected into embryonal carcinoma stem cells are both lost and inactivated at high frequency

Embryonal carcinoma (EC) cells can be efficiently transfected with cloned DNAs but there is a strong tendency for expression from transfected genes to be lost from stably transformed cells. To investigate the mechanism responsible for this loss of expression, we transfected P19 EC cells with a gene encoding the E. coli beta-galactosidase and examined expression of this gene in clonal populations of cells. Cells that carry and express the beta-galactosidase gene give rise to cells that do not express at a rate of about 0.02 events per cell per cell division. These non-expressing cells were of two types, some had lost the transfected genes while others had inactivated them. In those cells that retained but inactivated the transfected genes, the inactive state was stable and suppression was at the level of transcription initiation but not associated with increased DNA methylation. Because transfected DNAs integrate into the genome as tandem arrays, the gene loss and inactivation seen in EC cells may be analogous to the repeat-induced gene inactivation seen in lower eukaryotes.

The transcriptionally competent U2 gene is necessary and sufficient for adenovirus type 12 induction of the fragile site at 17q21-22

Adenovirus type 12 induces four fragile sites upon infection of human cells. The U2 locus, consisting of up to 20 tandem repeats of a 5.8-kbp monomer, maps at the most sensitive of these sites at 17q21-22. We have previously shown that an artificial U2 locus integrated into the human genome generates a new virus-induced fragile site. To determine which elements within the U2 monomer are responsible for fragility, we constructed loci consisting of tandem repeats of subfragments of the U2 monomer. With this approach, we demonstrate that a transcriptionally competent U2 gene is necessary and sufficient for virus-induced fragility and that no other element within the 5.8-kbp monomer contributes to this effect.