Homologous recombination between transfected DNAs.

An extensive analysis of the fate and structure of polyomavirus-plasmid recombinant molecules transfected into Rat-1 cells has revealed that the DNA often becomes integrated within transformed cell DNA in a head-to-tail tandem arrangement. This occurs independently of the replicative capacity of the transforming DNA and is facilitated by the use of large quantities of DNA during transfection. These observations have led us to suggest that head-to-tail tandems are formed by homologous recombination between transfected DNAs either before or after integration within cellular DNA. To test this hypothesis, we have measured the transforming activity of pairs of mutant, nontransforming, recombinant plasmid DNAs that carry different lesions in the transforming gene of polyomavirus. The results show that, although the individual mutant DNAs are incapable of transformation, transfection with pairs of mutant DNAs leads to the formation of transformed cells at high frequency. Moreover, there is a direct relationship between the distance between the lesions in pairs of mutant DNAs and their transforming activity. Finally, analyses of the structures of integrated recombinant plasmid DNAs and the viral proteins within independent transformed cells prove that recombination occurs between the mutant genomes to generate a wild-type transforming gene.     

Arrangement of Integrated Avian Sarcoma Virus DNA Sequences Within the Cellular Genomes of Transformed and Revertant Mammalian Cells

We have examined the arrangement of integrated avian sarcoma virus (ASV) DNA sequences in several different avian sarcoma virus transformed mammalian cell lines, in independently isolated clones of avian sarcoma virus transformed rat liver cells, and in morphologically normal revertants of avian sarcoma virus transformed rat embryo cells. By using restriction endonuclease digestion, agarose gel electrophoresis, Southern blotting, and hybridization with labeled avian sarcoma virus complementary DNA probes, we have compared the restriction enzyme cleavage maps of integrated viral DNA and adjacent cellular DNA sequences in four different mouse and rat cell lines transformed with either Bratislava 77 or Schmidt-Ruppin strains of avian sarcoma virus. The results of these experiments indicated that the integrated viral DNA resided at a different site within the host cell genome in each transformed cell line. A similar analysis of several independently derived clones of Schmidt-Ruppin transformed rat liver cells also revealed that each clone contained a unique cellular site for the integration of proviral DNA. Examination of several morphologically normal revertants and spontaneous retransformants of Schmidt-Ruppin transformed rat embryo cells revealed that the internal arrangement and cellular integration site of viral DNA sequences was identical with that of the transformed parent cell line. The loss of the transformed phenotype in these revertant cell lines, therefore, does not appear to be the result of rearrangement or deletions either within the viral genome or in adjacent cellular DNA sequences. The data presented support a model for ASV proviral DNA integration in which recombination can occur at multiple sites within the mammalian cell genome. The integration and maintenance of at least one complete copy of the viral genome appear to be required for continuous expression of the transformed phenotype in mammalian cells.     

Characterization of flat revertant cells isolated from simian virus 40-transformed mouse and rat cells which contain multiple copies of viral genomes.

Eight clones of flat revertants were isolated by negative selection from simian virus 40 (SV40)-transformed mouse and rat cell lines in which two and six viral genome equivalents per cell were integrated, respectively. These revertants showed either a normal cell phenotype or a phenotype intermediate between normal and transformed cells as to cellular morphology and saturation density and were unable to grow in soft agar medium. One revertant derived from SV40-transformed mouse cells was T antigen positive, whereas the other seven revertants were T antigen negative. SV40 could be rescued only from the T-antigen-positive revertant by fusion with permissive monkey cells. The susceptibility of the revertants to retransformation by wild-type SV40 was variable among these revertants. T-antigen-negative revertants from SV40-transformed mouse cells were retransformed at a frequency of 3 to 10 times higher than their grandparental untransformed cells. In contrast, T-antigen-negative revertants from SV40-transformed rat cells could not be retransformed. The arrangement of viral genomes was analyzed by digestion of cellular DNA with restriction enzymes of different specificity, followed by detection of DNA fragments containing a viral sequence and rat cells were serially arranged within the length of about 30 kilobases, with at least two intervening cellular sequences. A head-to-tail tandem array of unit length viral genomes was present in at least one insertion site in the transformed rat cells. All of the revertants had undergone a deletion(s), and only a part of the viral genome was retained in T-antigen-negative revertants. A relatively high frequency of reversion in the transformed rat cells suggests that reversion occurs by homologous recombination between the integrated viral genomes.     

Integration sites and sequence arrangement of SV40 DNA in a homogeneous series of transformed rat fibroblast lines

The state and organization of viral DNA sequences present in independently isolated rat cell lines transformed with SV40 were investigated using restriction-enzyme cleavage of the cellular DNA and blot hybridization with a viral probe. The transformed lines were established under conditions as identical as possible, except for a limited number of variables (multiplicity of infection, physiological state of the cells after infection and procedures used for selecting the transformed derivatives). They were characterized after a limited number of generations in culture. Two distinct types of organization were found: covalently integrated viral genomes were present either as single inserts or as head-to-tail oligomeric structures. The latter was observed among transformants derived from cells maintained after infection under growth-inhibiting conditions (suspension in agarose medium, confluency on a solid substrate). Single inserts were observed only among cell lines isolated after an initial period of active growth. Recurrent patterns of hybridizaton were observed in independently isolated lines, indicating that the sites of the integrative recombinations were close enough, both in the viral and the cellular sequences, not to be distinguished at the level of sensitivity of the technique (more than +/- 100 bp). Among cell lines with multiple integration sites, only part of the inserts were found in several instances to be identical to inserts observed in other transformed lines.     

Rescue of silent integrated polyoma genomes suggests homologous recombination between resident and transfected DNA fragments

Two defective polyoma virus genomes, deleted in the nucleotide sequences coding the N-termini of the tumor antigens, were introduced into Fisher 3T3 rat cells by DNA-mediated gene transfer (transfection). The resulting integrated genomes were incapable of conferring a transformed phenotype to the cells. However, after transfection of these lines with small polyoma fragments overlapping the deleted sequences, transformed clones were isolated. These clones were analyzed by Southern genomic blot hybridization and by isolation in E. coli of plasmids containing viral sequences excised following fusion with mouse polyoma growth-permissive cells. In all cases at least one intact copy of the early region of the polyoma genome was found. Furthermore, restriction sites adjacent to the initial inactive insertion remained unchanged in many of the transformed lines. These results show that functional restoration of the defective polyoma early region involves homologous recombination between the deleted viral genomes integrated in the cellular DNA and the transfecting viral fragments.     

Low probability of double integration in transformation of nonpermissive cells by polyomavirus.

The fate of polyomavirus genomes in stable transformants was examined in experiments in which rat or hamster cells were infected with mixed viral populations containing either two distinguishable wild types or a wild type plus a transformation-deficient mutant. The results demonstrate that in 90% of the cases in either situation, a single polyomavirus parental genome became integrated into the host genome. Both parents in the mixed infection were present in the same cells and both persisted in the infected cells until transformants appeared, eliminating the possibility that one virus would exclude the other in the early steps of the infection process. We believe that these results can be generalized to transformation events derived from normal single infections. Thus, contrary to previous results for the number of integration sites determined by restriction endonuclease and blot hybridization analysis, it appears that the probability for more than one integration event in transformation by polyomavirus is low (1 in 10). A reinterpretation of previous data is proposed.     

Integration and expression of viral DNA in cells transformed by host range mutants of adenovirus type 5.

Group I host range (hr) mutants of adenovirus type 5 are unable to transform rat embryo or rat embryo brain cells but induce an abnormal transformation of baby rat kidney cells. We established several transformed rat kidney cell lines and characterized them with respect to the transformed phenotype and the structure of the integrated viral DNA. The hr mutant-transformed cells, unlike wild-type virus transformants, were fibroblastic rather than epithelial, failed to grow in soft agar, and were also less tumorigenic in nude mice. Studies on the structure of the integrated viral DNA sequences showed that hr-transformed cells always contained the left end of the adenovirus DNA, but the size of the integrated DNA fragment varied among different lines, and a high percentage of the lines contained the entire viral genome colinearly integrated. The patterns of integration were maintained after prolonged growth in culture and after subcloning. Attempts to rescue infectious virus from lines which contained the entire genome were unsuccessful. Using immunoprecipitation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis, we analyzed the viral proteins expressed in hr-transformed cells. Results of these studies indicated that, like wild type-transformed cells, hr transformants expressed E1B proteins of molecular weight 58,000 and 19,000.     

Analysis of stable and unstable viral forms in SV40-infected human keratinocytes

We analyzed the state of the genomic DNA of the papovavirus SV40 in human keratinocytes as viral-infected cells gradually acquired a transformed phenotype over time. Initially, the vast majority of the viral DNA is maintained either in a full-length supercoiled form or as truncated subgenomic fragments with little evidence of integration. However, analyses of clonal populations revealed great heterogeneity and instability of the viral DNA, and we were able to isolate one clonal subpopulation in which integrated forms of the virus appeared to predominate. Similarly, uncloned populations eventually ceased production of the "free" viral DNA after several years in culture and instead came to display tandemly repeated SV40 copies at a single host integration site. Interestingly, Bg1 II digestion of host DNA generated restriction fragments containing the integrated SV40 DNA, which were of differing sizes in cultures at the 144th vs the 163rd serial passage suggesting modification or rearrangement of sequences at or near the integration site. Host sequences flanking the integrated viral DNA at the 163rd serial passage have been isolated on restriction fragments generated by Eco RI, Bam HI, and Hpa II digestion. These analyses suggest that the integrated virus is linearized near the Bg1 I site and contains a large deletion in the SV40 early region at one of the viral-host junctions.     

Processing of DNA prior to illegitimate recombination in mouse cells.

In mammalian cells, the predominant pathway of chromosomal integration of exogenous DNA is random or illegitimate recombination; integration by homologous recombination is infrequent. Homologous recombination is initiated at double-strand DNA breaks which have been acted on by single-strand exonuclease. To further characterize the relationship between illegitimate and homologous recombination, we have investigated whether illegitimate recombination is also preceded by exonuclease digestion. Heteroduplex DNAs which included strand-specific restriction markers at each of four positions were generated. These DNAs were introduced into mouse embryonic stem cells, and stably transformed clones were isolated and analyzed to determine whether there was any strand bias in the retention of restriction markers with respect to their positions. Some of the mismatches appear to have been resolved by mismatch repair. Very significant strand bias was observed in the retention of restriction markers, and there was polarity of marker retention between adjacent positions. We conclude that DNA is frequently subjected to 5'-->3' exonuclease digestion prior to integration by illegitimate recombination and that the length of DNA removed by exonuclease digestion can be extensive. We also provide evidence which suggests that frequent but less extensive 3'-->5' exonuclease processing also occurs.     

Recombination between bacteriophage T4 and plasmid pBR322 molecules containing cloned T4 DNA

Reciprocal recombination between T4 DNA cloned in plasmid pBR322 and homologous sequences in bacteriophage T4 genomes leads to integration of complete plasmid molecules into phage genomes. Indirect evidence of this integration comes from two kinds of experiments. Packaging of pBR322 DNA into mature phage particles can be detected by a DNA--DNA hybridization assay only when a T4 restriction fragment is cloned in the plasmid. The density of the pBR322 DNA synthesized after phage infection is also consistent with integration of plasmid vector DNA into vegetative phage genomes. Direct evidence of plasmid integration into phage genomes in the region of DNA homology comes from genetic and biochemical analysis of cytosine-containing DNA isolated from mature phage particles. Agarose gel electrophoresis of restriction endonuclease-digested DNA, followed by Southern blot analysis with nick-translated probes, shows that entire plasmid molecules become integrated into phage genomes in the region of T4 DNA homology. In addition, this analysis shows that genomes containing multiple copies of complete plasmid molecules are also formed. Among phage particles containing at least one integrated copy, the average number of integrated plasmid molecules is almost ten. A cloning experiment done with restricted DNA confirms these conclusions and illustrates a method for walking along the T4 genome.     

Constraints to genetic exchange support gene coadaptation in a tripartite RNA virus

Genetic exchange by recombination, or reassortment of genomic segments, has been shown to be an important process in RNA virus evolution, resulting often in important phenotypic changes affecting host range and virulence. However, data from numerous systems indicate that reassortant or recombinant genotypes could be selected against in virus populations and suggest that there is coadaptation among viral genes. Little is known about the factors affecting the frequency of reassortants and recombinants along the virus life cycle. We have explored this issue by estimating the frequency of reassortant and recombinant genotypes in experimental populations of Cucumber mosaic virus derived from mixed infections with four different pairs of isolates that differed in about 12% of their nucleotide sequence. Genetic composition of progeny populations were analyzed at various steps of the virus life cycle during host colonization: infection of leaf cells, cell-to-cell movement within the inoculated leaf, encapsidation of progeny genomes, and systemic movement to upper noninoculated leaves. Results indicated that reassortant frequencies do not correspond to random expectations and that selection operates against reassortant genotypes. The intensity of selection, estimated through the use of log-linear models, increased as host colonization progressed. No recombinant was detected in any progeny. Hence, results showed the existence of constraints to genetic exchange linked to various steps of the virus life cycle, so that genotypes with heterologous gene combinations were less fit and disappeared from the population. These results contribute to explain the low frequency of recombinants and reassortants in natural populations of many viruses, in spite of high rates of genetic exchange. More generally, the present work supports the hypothesis of coadaptation of gene complexes within the viral genomes.     

Patterns of integration of viral dna sequences in the genomes of adenovirus type 12-transformed hamster cells

The patterns of integration of the viral genome have been analyzed in four hamster cell lines transformed by adenovirus type 12 (Ad12). It has previously been shown that in each of the cell lines HA12/7, T637, A2497-2 and A2497-3, the viral genome persists in multiple copies, and that different parts of the viral DNA are represented non-stoichiometrically (Fanning and Doerfler, 1976). All four cell lines are oncogenic when injected into hamsters.The DNA from each of the cell lines was extracted and cleaved in different experiments with restriction endonucleases Bam HI, Bgl II, Eco RI, Hind III, Hpa II or Sma I. The DNA fragments were separated on 1% agarose slab gels and transferred to nitrocellulose filters by the Southern technique. Ad12 DNA sequences were detected by hybridization to Ad12 DNA, which was ³²P-labeled by nick translation, and by subsequent autoradiography. In some experiments, the ³²P-labeled Eco RI restriction endonuclease fragments of Ad12 DNA were used to investigate the distribution of specific segments of the viral genome in the cellular DNA.For each cell line, a distinct and specific pattern of integrated viral DNA sequences is observed for each of the restriction endonucleases used. Moreover, viral sequences complementary to the isolated Eco RI restriction endonuclease fragments are also distributed in patterns specific for each cell line. There are striking differences in integration patterns among the four different lines; there are also similarities. Because the organization of cellular genes in virus-transformed as compared to normal cells has not yet been determined, conclusions about the existence or absence of specific integration sites for adenovirus DNA appear premature. Analysis of the integration patterns of Ad12 DNA in the four hamster lines investigated reveals that some of the viral DNA molecules are fragmented prior to or during integration. Analysis with specific restriction endonuclease fragments demonstrates that the Eco RI B, D and E fragments, comprising a contiguous segment from 0.17–0.62 fractional length units of the viral DNA, remain intact during integration in a portion of the viral DNA molecules. Although each cell line carries multiple copies of Ad12 DNA, the viral DNA sequences are concentrated in a small number of distinct size classes of fragments. This finding is compatible with, but does not prove, the notion that at least a portion of the viral DNA sequences is integrated into repetitive sequences, or else that the integrated viral sequences have been amplified after integration.In the three cell lines which were tested, the integration pattern is stable over many generations, with continuous passage-twice weekly-of cells for 6–7 months. In the three cell lines which were examined, the integration pattern is identical in a number of randomly isolated clones. Hence it can be concluded that the patterns of integration are identical among all cells in a population of a given line of transformed cells.     

Homologous recombination and double-strand break repair in the transformation of Rhizopus oryzae

Genetic transformation of the Mucorales fungi has been problematic, since DNA transformed into the host rarely integrates and usually is mitotically unstable in the absence of selective pressure. In this study, transformation of Rhizopus oryzae was investigated to determine if the fate of introduced DNA could be predicted based on double-strand break repair and recombination mechanisms found in other fungi. A transformation system was developed with uracil auxotrophs of Rhizopus oryzae that could be complemented with the pyrG gene isolated in this work. DNA transformed as circular plasmids was maintained extrachromosomally in high-molecular-weight (>23 kb) concatenated arrangement. Type-I crossover integration into the pyrG locus and type-III pyrG gene replacement events occurred in approximately 1-5% of transformants. Linearization of the plasmid pPyr225 with a single restriction enzyme that cleaves within the vector sequence almost always resulted in isolates with replicating concatenated plasmids that had been repaired by end-joining recombination that restored the restriction site. The addition of a 40-bp direct repeat on either side of this cleavage site led to repair by homologous recombination between the repeated sequences on the plasmid, resulting in loss of the restriction site. When plasmid pPyr225 was digested with two different enzymes that cleave within the vector sequence to release the pyrG containing fragment, only pyrG gene replacement recombination occurred in transformants. Linearization of plasmid pPyr225 within the pyrG gene itself gave the highest percentage (20%) of type-I integration at the pyrG locus. However, end-joining repair and gene replacement events were still the predominant types of recombination found in transformations with this plasmid topology.     

'Integron'-bearing vectors: a method suitable for stable chromosomal integration in highly restrictive corynebacteria.

A pBR322-derived plasmid (pCGL107) that carries the Corynebacterium melassecola ATCC17965 analogue of Escherichia coli gdhA gene (encoding glutamate dehydrogenase), was introduced into the related strain, Brevibacterium lactofermentum CGL2002, by electroporation and integrated into its chromosome by homologous recombination. However, pCGL107 cannot integrate into C. melassecola, since the host restriction prevents successful electroporation by E. coli-modified DNA. Nevertheless, B. lactofermentum-modified replicative plasmid DNA can be transformed by electroporation into C. melassecola; thus pCGL519-2, a shuttle plasmid that carries the C. melassecola analogue of E. coli gltA (encoding citrate synthase), was extracted from the former host and electroporated into the latter. Rare restriction sites conveniently placed in pCGL519-2 were used to recover a replicon-less cartridge called 'integron', that contains a selectable marker and gtlA within a single fragment. Integron prepared from pCGL519-2 DNA which had been extracted from C. melassecola, and thus, was capable of eluding the C. melassecola restriction barrier(s), was successfully electroporated into this host. The molecular analysis of the resulting transformants suggests that they result from the integration of a single circular integron molecule by homologous recombination between the gltA regions of the host genome and the integron. These transformants were stable for 30 generations in the absence of selection.     

Analysis of integrated avian RNA tumor virus DNA in transformed chicken, duck and quail fibroblasts.

The state of integration of avian sarcoma virus DNA in the genomes of transformed chicken, duck, and quail fibroblasts was deduced by means of restriction enzyme digestion of total cell DNA, gel electrophoresis, and subsequent analysis by the procedure of Southern. The cells used in these studies were either mass-infected cultures or clones of infected cells selected by their ability to form colonies in agar. For both mass-infected cultures and clones of cells of all three species, we found that integration occurred at a specific site on the viral genome but appeared to occur at many sites on the cell genome. At least some of the integrated viral DNA existed as intact nonpermuted species flanked by direct terminal repeats of at least 0.134 megadalton (217 base pairs). For each of 12 transformed quail clones studied, it was possible to detect, after digestion with Kpn I, unique junctions between viral and cellular DNA. That is, at our level of analysis, the integration site on the cell genome for each clone was different. However, within each of the 17 chicken and 9 duck clones of transformed cells, a heterogeneity presumably occurred during the outgrowth of the cell clone population, in that we could not readily detect identifiable cell-virus junction fragments.     

Induction of recombination between homologous and diverged DNAs by double-strand gaps and breaks and role of mismatch repair.

Sequence homology is expected to influence recombination. To further understand mechanisms of recombination and the impact of reduced homology, we examined recombination during transformation between plasmid-borne DNA flanking a double-strand break (DSB) or gap and its chromosomal homolog. Previous reports have concentrated on spontaneous recombination or initiation by undefined lesions. Sequence divergence of approximately 16% reduced transformation frequencies by at least 10-fold. Gene conversion patterns associated with double-strand gap repair of episomal plasmids or with plasmid integration were analyzed by restriction endonuclease mapping and DNA sequencing. For episomal plasmids carrying homeologous DNA, at least one input end was always preserved beyond 10 bp, whereas for plasmids carrying homologous DNA, both input ends were converted beyond 80 bp in 60% of the transformants. The system allowed the recovery of transformants carrying mixtures of recombinant molecules that might arise if heteroduplex DNA--a presumed recombination intermediate--escapes mismatch repair. Gene conversion involving homologous DNAs frequently involved DNA mismatch repair, directed to a broken strand. A mutation in the PMS1 mismatch repair gene significantly increased the fraction of transformants carrying a mixture of plasmids for homologous DNAs, indicating that PMS1 can participate in DSB-initiated recombination. Since nearly all transformants involving homeologous DNAs carried a single recombinant plasmid in both Pms+ and Pms- strains, stable heteroduplex DNA appears less likely than for homologous DNAs. Regardless of homology, gene conversion does not appear to occur by nucleolytic expansion of a DSB to a gap prior to recombination. The results with homeologous DNAs are consistent with a recombinational repair model that we propose does not require the formation of stable heteroduplex DNA but instead involves other homology-dependent interactions that allow recombination-dependent DNA synthesis.     

Linear DNA must have free ends to transform rat cells efficiently

Summary We have observed that failure to remove certain restriction enzymes after digestion reduced the transforming ability of DNA from 10- to 50-fold. The DNA found integrated in the transformed cells isolated under these conditions had lost little or no sequences. We interpret these results as indicating that certain restriction enzymes remain bound to the DNA ends after digestion, thus generating a substrate unfavorable both for integration and exonucleolytic degradation. As expected from this interpretation, removal of the restriction enzymes before transfection restored the full transforming ability of linear DNA, but also resulted in the integrated sequences being significantly shorter than the transfected DNA. These findings strongly argue for the hypothesis that integration of linear DNA by illegitimate recombination requires free ends and further suggest that exonucleolytic degradation of such ends may generate a preferred substrate for integration. Finally, a comparison of the sequences found integrated after transfection with circular or linear molecules, led us to conclude that circular molecules need not be linearized to become integrated.     

M-MuLV-induced leukemogenesis: Integration and structure of recombinant proviruses in tumors

M-MuLV-specific DNA probes were used to establish the state of integration and amplification of recombinant proviral sequences in Moloney virus-induced tumors of Balb/Mo, Balb/c and 129 mice. The somatically acquired viral sequences contain both authentic M-MuLV genomes and recombinants of M-MuLV with endogenous viral sequences. All reintegrated genomes carry long terminal repeat (LTR) sequences at both termini of their genome. In the preleukemic stage a large population of cells exhibiting a random distribution of reintegrated M-MuLV genomes are seen, but during outgrowth of the tumor, selection of cells occurs leaving one or a few clonal descendants in the outgrown tumor. In this latter stage recombinant genomes can be detected. Although these recombinants constitute a heterogeneous group of proviruses, characteristic molecular markers are conserved among many individual proviral recombinants, lending credence to the notion that a certain recombinant structure is a prerequisite for the onset of neoplasia. The structure of these recombinants shows close structural similarities to the previously described mink cell focus-inducing (MCF)-type viruses.     

Restriction enzyme-mediated integration used to produce pathogenicity mutants of Colletotrichum graminicola

We have developed a restriction enzyme-mediated insertional mutagenesis (REMI) system for the maize pathogen Colletotrichum graminicola. In this report, we demonstrate the utility of a REMI-based mutagenesis approach to identify novel pathogenicity genes. Use of REMI increased transformation efficiency by as much as 27-fold over transformations with linearized plasmid alone. Ninety-nine transformants were examined by Southern analysis, and 51% contained simple integrations consisting of one copy of the vector integrated at a single site in the genome. All appeared to have a plasmid integration at a unique site. Sequencing across the integration sites of six transformants demonstrated that in all cases the plasmid integration occurred at the corresponding restriction enzyme-recognition site. We used an in vitro bioassay to identify two pathogenicity mutants among 660 transformants. Genomic DNA flanking the plasmid integration sites was used to identify corresponding cosmids in a wild-type genomic library. The pathogenicity of one of the mutants was restored when it was transformed with the cosmids.     

Monitoring Insect Transposable Elements in Large Double-Stranded DNA Viruses Reveals Host-to-Virus and Virus-to-Virus Transposition

The mechanisms by which transposable elements (TEs) can be horizontally transferred between animals are unknown, but viruses are possible candidate vectors. Here, we surveyed the presence of host-derived TEs in viral genomes in 35 deep sequencing data sets produced from 11 host–virus systems, encompassing nine arthropod host species (five lepidopterans, two dipterans, and two crustaceans) and six different double-stranded (ds) DNA viruses (four baculoviruses and two iridoviruses). We found evidence of viral-borne TEs in 14 data sets, with frequencies of viral genomes carrying a TE ranging from 0.01% to 26.33% for baculoviruses and from 0.45% to 7.36% for iridoviruses. The analysis of viral populations separated by a single replication cycle revealed that viral-borne TEs originating from an initial host species can be retrieved after viral replication in another host species, sometimes at higher frequencies. Furthermore, we detected a strong increase in the number of integrations in a viral population for a TE absent from the hosts’ genomes, indicating that this TE has undergone intense transposition within the viral population. Finally, we provide evidence that many TEs found integrated in viral genomes (15/41) have been horizontally transferred in insects. Altogether, our results indicate that multiple large dsDNA viruses have the capacity to shuttle TEs in insects and they underline the potential of viruses to act as vectors of horizontal transfer of TEs. Furthermore, the finding that TEs can transpose between viral genomes of a viral species sets viruses as possible new niches in which TEs can persist and evolve.