Acceptor sites for retroviral integrations map near DNase I-hypersensitive sites in chromatin.

Seven cellular loci with acceptor sites for retroviral integrations have been mapped for the presence of DNase I-hypersensitive sites in chromatin. Integrations in three of these loci, chicken c-erbB, rat c-myc, and a rat locus, dsi-1, had been selected for in retrovirus-induced tumors. Of the remaining four, two, designated dsi-3 and dsi-4, harbored acceptor sites for apparently unselected integrations of Moloney murine leukemia virus in a Moloney murine leukemia virus-induced thymoma, and two, designated C and F, harbored unselected acceptor sites for Moloney murine leukemia virus integrations in a rat fibroblast cell line. Each acceptor site mapped to within 500 base pairs of a DNase I-hypersensitive site. In the analyses of the unselected integrations, six hypersensitive sites were observed in 39 kilobases of DNA. The four acceptor sites in this DNA were localized between 0.05 and 0.43 kilobases of a hypersensitive site. The probability of this close association occurring by chance was calculated to be extremely low. Hypersensitive sites were mapped in cells representing the lineage in which integration had occurred as well as in an unrelated lineage. In six of the seven acceptor loci hypersensitive sites could not be detected in the unrelated lineage. Our results indicate that retroviruses preferentially integrate close to DNase I-hypersensitive sites and that many of these sites are expressed in some but not all cells.     

Retroviral integration into minichromosomes in vitro.

We describe here the use of chromatin as a target for retroviral integration in vitro. Extracts of cells newly infected with murine leukemia virus (MLV) provided the source of integration activity, and yeast TRP1ARS1 and SV40 minichromosomes served as simple models for chromatin. Both minichromosomes were used as targets for integration, with efficiencies comparable with that of naked DNA. In addition, under some reaction conditions the minichromosomes behaved as if they were used preferentially over naked DNAs in the same reaction. Mapping of integration sites by cloning and sequencing recombinants revealed that the integration machinery does not display a preference for nucleosome-free, nuclease-sensitive regions. The distributions of integration sites in TRP1ARS1 minichromosomes and a naked DNA counterpart were grossly similar, but in a detailed analysis the distribution in minichromosomes was found to be significantly more ordered: the sites displayed a periodic spacing of approximately 10 bp, many sites sustained multiple insertions and there was sequence bias at the target sites. These results are in accord with a model in which the integration machinery has preferential access to the exposed face of the nucleosomal DNA helix. The population of potential sites in chromatin therefore becomes more limited, in a manner dictated by the rotational orientation of the DNA sequence around the nucleosome core, and those sites are used more frequently than in naked DNA.     

Fidelity of Target Site Duplication and Sequence Preference during Integration of Xenotropic Murine Leukemia Virus-Related Virus

Xenotropic murine leukemia virus (MLV)-related virus (XMRV) is a new human retrovirus associated with prostate cancer and chronic fatigue syndrome. The causal relationship of XMRV infection to human disease and the mechanism of pathogenicity have not been established. During retrovirus replication, integration of the cDNA copy of the viral RNA genome into the host cell chromosome is an essential step and involves coordinated joining of the two ends of the linear viral DNA into staggered sites on target DNA. Correct integration produces proviruses that are flanked by a short direct repeat, which varies from 4 to 6 bp among the retroviruses but is invariant for each particular retrovirus. Uncoordinated joining of the two viral DNA ends into target DNA can cause insertions, deletions, or other genomic alterations at the integration site. To determine the fidelity of XMRV integration, cells infected with XMRV were clonally expanded and DNA sequences at the viral-host DNA junctions were determined and analyzed. We found that a majority of the provirus ends were correctly processed and flanked by a 4-bp direct repeat of host DNA. A weak consensus sequence was also detected at the XMRV integration sites. We conclude that integration of XMRV DNA involves a coordinated joining of two viral DNA ends that are spaced 4 bp apart on the target DNA and proceeds with high fidelity.     

Mink cell focus-forming murine leukemia virus infection induces apoptosis of thymic lymphocytes

In a previous study we identified the subpopulations of thymus cells that were infected by the lymphomagenic MCF13 murine leukemia virus (MLV) (F. K. Yoshimura, T. Wang, and M. Cankovic, J. Virol. 73:4890-4898, 1999) and observed an effect on thymus size by virus infection. In this report we describe our results which demonstrate that MCF13 MLV infection of thymuses reduced the number of T lymphocytes in this organ. Histological examination showed diffuse lymphocyte depletion, which was most striking in the CD4(+) CD8(+) lymphocyte-enriched cortical zone. Consistent with this, flow cytometric analysis showed that the lymphocytes which were depleted were predominantly the immature CD3(-) CD4(+) CD8(+) and CD3(+) CD4(+) CD8(+) cells. A comparison of the percentages of live, apoptotic, and dead cells of the gp70(+) and gp70(-) thymic lymphocytes suggested that this effect on thymus cellularity is a result of virus infection. Studies of the survival of thymic T lymphocytes in culture showed that cells from MCF13 MLV-inoculated mice underwent greater apoptosis and death than cells from control animals. Assays for apoptosis included 7-amino-actinomycin D staining, DNA fragmentation, and cleavage of caspase-3 and poly(ADP-ribose) polymerase proenzymes. Our results suggest that apoptosis of thymic lymphocytes by virus infection is an important step in the early stages of MCF13 MLV tumorigenesis.     

DNA bending creates favored sites for retroviral integration: an explanation for preferred insertion sites in nucleosomes.

The choice of retroviral integration sites is strongly influenced by chromatin: integration in vitro occurs more efficiently into nucleosomal DNA than into naked DNA, and a characteristic pattern of preferred insertion sites with a 10 bp periodicity is observed at the outer face of the nucleosomal DNA. At least three features of nucleosomal DNA could be responsible for the creation of these favored sites: the presence of histones, attachment of the DNA to a protein surface, and DNA bending. To test each of these possibilities, we studied integration in vitro with human immunodeficiency virus and murine leukemia virus integrases into four model targets that mimic features of nucleosomal DNA: (i) catabolite activator protein-DNA complexes; (ii) lac repressor-operator complexes; (iii) lac repressor-induced loops; and (iv) intrinsically bent A-tract DNA. We found that bending of the target DNA can create favored integration sites at the outer face of the helix, irrespective of whether the bent DNA is attached to a protein surface. Our findings offer an explanation for the preferred usage of nucleosomes as integration targets. In addition, they suggest that bending of the target DNA might be an intrinsic feature of the integration reaction.     

Methylation state and DNase I sensitivity of chromatin containing Moloney murine leukemia virus DNA in exogenously infected mouse cells

The nature of Moloney murine leukemia virus (M-MuLV)-specific proviral DNA in exogenously infected mouse cells was studied. M-MuLV clone A9 cells, NIH-3T3 fibroblasts productively infected with M-MuLV, were used. These cells contain 10 to 15 copies of M-MuLV proviral DNA. The state of methylation of M-MuLV proviral DNA was examined by cleaving A9 cell DNA with restriction endonucleases which have the dinucleotide CpG in their cleavage sequences. Analysis with such enzymes, which recognized nine different sites in M-MuLV DNA, indicated that most if not all of the M-MuLV proviruses in A9 cells were completely unmethylated. An individual proviral integration was examined, using as probe adjacent single-copy cellular sequences. These sequences were obtained from a lambda phage recombinant clone containing an M-MuLV provirus from the A9 cells. This individual integration also showed no detectable methylation. In contrast, endogenous MuLV-related sequences present in NIH-3T3 cells before infection were largely methylated. The configuration chromatin containing M-MuLV proviruses was also investigated by digesting A9 nuclei with DNase I, followed by restriction analysis of the remaining DNA. Endogenous MuLV-related DNA was in chromatin relatively resistant to DNase I digestion, whereas the majority of M-MuLV-specific proviruses were in domains of intermediate DNase I sensitivity. Two proviral copies hypersensitive to DNase I digestion were identified. Analogy to the DNase I sensitivity of expressed and nonexpressed globin genes suggested that the proviral copies containing DNase I-hypersensitive sites were transcribed.     

Deciphering the code for retroviral integration target site selection

Upon cell invasion, retroviruses generate a DNA copy of their RNA genome and integrate retroviral cDNA within host chromosomal DNA. Integration occurs throughout the host cell genome, but target site selection is not random. Each subgroup of retrovirus is distinguished from the others by attraction to particular features on chromosomes. Despite extensive efforts to identify host factors that interact with retrovirion components or chromosome features predictive of integration, little is known about how integration sites are selected. We attempted to identify markers predictive of retroviral integration by exploiting Precision-Recall methods for extracting information from highly skewed datasets to derive robust and discriminating measures of association. ChIPSeq datasets for more than 60 factors were compared with 14 retroviral integration datasets. When compared with MLV, PERV or XMRV integration sites, strong association was observed with STAT1, acetylation of H3 and H4 at several positions, and methylation of H2AZ, H3K4, and K9. By combining peaks from ChIPSeq datasets, a supermarker was identified that localized within 2 kB of 75% of MLV proviruses and detected differences in integration preferences among different cell types. The supermarker predicted the likelihood of integration within specific chromosomal regions in a cell-type specific manner, yielding probabilities for integration into proto-oncogene LMO2 identical to experimentally determined values. The supermarker thus identifies chromosomal features highly favored for retroviral integration, provides clues to the mechanism by which retrovirus integration sites are selected, and offers a tool for predicting cell-type specific proto-oncogene activation by retroviruses.     

In vitro murine leukemia retroviral integration and structure fluctuation of target DNA

Integration of the retroviral genome into host DNA is a critical step in the life cycle of a retrovirus. Although assays for in vitro integration have been developed, the actual DNA sequences targeted by murine leukemia retrovirus (MLV) during in vitro reproduction are unknown. While previous studies used artificial target sequences, we developed an assay using target DNA sequences from common MLV integration sites in Stat5a and c-myc in the genome of murine lymphomas and successfully integrated MLV into the target DNA in vitro. We calculated the free energy change during folding of the target sequence DNA and found a close correlation between the calculated free energy change and the number of integrations. Indeed, the integrations closely correlated with fluctuation of the structure of the target DNA segment. These data suggest that the fluctuation may generate a DNA structure favorable for in vitro integration into the target DNA. The approach described here can provide data on the biochemical properties of the integration reaction to which the target DNA structure may contribute.     

No evidence for XMRV nucleic acids, infectious virus or anti-XMRV antibodies in Canadian patients with chronic fatigue syndrome

The gammaretroviruses xenotropic murine leukemia virus (MLV)-related virus (XMRV) and MLV have been reported to be more prevalent in plasma and peripheral blood mononuclear cells of chronic fatigue syndrome (CFS) patients than in healthy controls. Here, we report the complex analysis of whole blood and plasma samples from 58 CFS patients and 57 controls from Canada for the presence of XMRV/MLV nucleic acids, infectious virus, and XMRV/MLV-specific antibodies. Multiple techniques were employed, including nested and qRT-PCR, cell culture, and immunoblotting. We found no evidence of XMRV or MLV in humans and conclude that CFS is not associated with these gammaretroviruses.     

Primate gammaretroviruses require an ancillary factor not required for murine gammaretroviruses to infect BHK cells

BHK cells remain resistant to xenotropic murine retrovirus-related virus (XMRV) or gibbon ape leukemia virus (GALV) infection, even when their respective receptors, Xpr1 or PiT1, are expressed. We set out to determine the stage at which viral infection is blocked and whether this block is mediated by a dominant-negative factor or the absence of a requisite ancillary factor. BHK cells bind neither XMRV nor GALV envelope proteins. BHK cells expressing the appropriate receptors bind XMRV or GALV envelope proteins. BHK cells can be infected by NZB-XMV(New Zealand Black mouse xenotropic murine virus)-enveloped vectors, expressing an envelope derived from a xenotropic retrovirus that, like XMRV, employs Xpr1 as a receptor, and also by vectors bearing the envelope of 10A1 murine leukemia virus (MLV), a murine retrovirus that can use PiT1 as a receptor. The retroviral vectors used in these analyses differ solely in their viral envelope proteins, suggesting that the block to XMRV and GALV infection is mediated at the level of envelope-receptor interactions. N-linked glycosylation of the receptors was not found to mediate resistance of receptor-expressing BHK cells to GALV or XMRV, as shown by tunicamycin treatment and mutation of the specific glycosylation site of the PiT1 receptor. Hybrid cells produced by fusing BHKXpr1 or BHKPiT1 to XMRV- or GALV-resistant cells, respectively, can mediate efficient XMRV or GALV infection. These findings indicate that BHK cells lack a factor that is required for infection by primate xenotropic viruses. This factor is not required for viruses that use the same receptors but were directly isolated from mice.     

Absence of XMRV and closely related viruses in primary prostate cancer tissues used to derive the XMRV-infected cell line 22Rv1

The 22Rv1 cell line is widely used for prostate cancer research and other studies throughout the world. These cells were established from a human prostate tumor, CWR22, that was serially passaged in nude mice and selected for androgen independence. The 22Rv1 cells are known to produce high titers of xenotropic murine leukemia virus-related virus (XMRV). Recent studies suggested that XMRV was inadvertently created in the 1990's when two murine leukemia virus (MLV) genomes (pre-XMRV1 and pre-XMRV-2) recombined during passaging of the CWR22 tumor in mice. The conclusion that XMRV originated from mice and not the patient was based partly on the failure to detect XMRV in early CWR22 xenografts. While that deduction is certainly justified, we examined the possibility that a closely related virus could have been present in primary tumor tissue. Here we report that we have located the original prostate tumor tissue excised from patient CWR22 and have assayed the corresponding DNA by PCR and the tissue sections by fluorescence in situ hybridization for the presence of XMRV or a similar virus. The primary tumor tissues lacked mouse DNA as determined by PCR for intracisternal A type particle DNA, thus avoiding one of the limitations of studying xenografts. We show that neither XMRV nor a closely related virus was present in primary prostate tissue of patient CWR22. Our findings confirm and reinforce the conclusion that XMRV is a recombinant laboratory-generated mouse virus that is highly adapted for human prostate cancer cells.     

No association of xenotropic murine leukemia virus-related viruses with prostate cancer

BACKGROUND: The association of the xenotropic murine leukemia virus-related virus (XMRV) with prostate cancer continues to receive heightened attention as studies report discrepant XMRV prevalences ranging from zero up to 23%. It is unclear if differences in the diagnostic testing, disease severity, geography, or other factors account for the discordant results. We report here the prevalence of XMRV in a population with well-defined prostate cancers and RNase L polymorphism. We used broadly reactive PCR and Western blot (WB) assays to detect infection with XMRV and related murine leukemia viruses (MLV). METHODOLOGY/PRINCIPAL FINDINGS: We studied specimens from 162 US patients diagnosed with prostate cancer with a intermediate to advanced stage (Gleason Scores of 5-10; moderate (46%) poorly differentiated tumors (54%)). Prostate tissue DNA was tested by PCR assays that detect XMRV and MLV variants. To exclude contamination with mouse DNA, we also designed and used a mouse-specific DNA PCR test. Detailed phylogenetic analysis was used to infer evolutionary relationships. RNase L typing showed that 9.3% were homozygous (QQ) for the R462Q RNase L mutation, while 45.6% and 45.1% were homozygous or heterozygous, respectively. Serologic testing was performed by a WB test. Three of 162 (1.9%) prostate tissue DNA were PCR-positive for XMRV and had undetectable mouse DNA. None was homozygous for the QQ mutation. Plasma from all three persons was negative for viral RNA by RT-PCR. All 162 patients were WB negative. Phylogenetic analysis inferred a distinct XMRV. CONCLUSIONS AND THEIR SIGNIFICANCE: We found a very low prevalence of XMRV in prostate cancer patients. Infection was confirmed by phylogenetic analysis and absence of contaminating mouse DNA. The finding of undetectable antibodies and viremia in all three patients may reflect latent infection. Our results do not support an association of XMRV or MLV variants with prostate cancer.