Role of transposable elements in heterochromatin and epigenetic control

Heterochromatin has been defined as deeply staining chromosomal material that remains condensed in interphase, whereas euchromatin undergoes de-condensation. Heterochromatin is found near centromeres and telomeres, but interstitial sites of heterochromatin (knobs) are common in plant genomes and were first described in maize. These regions are repetitive and late-replicating. In Drosophila, heterochromatin influences gene expression, a heterochromatin phenomenon called position effect variegation. Similarities between position effect variegation in Drosophila and gene silencing in maize mediated by "controlling elements" (that is, transposable elements) led in part to the proposal that heterochromatin is composed of transposable elements, and that such elements scattered throughout the genome might regulate development. Using microarray analysis, we show that heterochromatin in Arabidopsis is determined by transposable elements and related tandem repeats, under the control of the chromatin remodelling ATPase DDM1 (Decrease in DNA Methylation 1). Small interfering RNAs (siRNAs) correspond to these sequences, suggesting a role in guiding DDM1. We also show that transposable elements can regulate genes epigenetically, but only when inserted within or very close to them. This probably accounts for the regulation by DDM1 and the DNA methyltransferase MET1 of the euchromatic, imprinted gene FWA, as its promoter is provided by transposable-element-derived tandem repeats that are associated with siRNAs.     

Similarity of the Cin1 repetitive family of Zea mays to eukaryotic transposable elements

It has been suggested that the middle repetitive class of sequences that make up a large proportion of the eukaryotic genome have been amplified and dispersed by DNA transposition. Transposition is a phenomenon first postulated by Barbara McClintock on the basis of her genetic analysis of mutants in Zea mays. Since then, DNA transposition has been studied genetically in various plant systems and is well documented on the molecular level in both prokaryotes and eukaryotes. This has included the isolation of DNA inserts at various loci in several plants; however, the prevalence of transposition in plants is not established. We report here DNA nucleotide sequence data which show that some members of the Cin1 middle repetitive family of maize have features characteristic of known transposable elements. One cloned Cin1 repeat has a 6-base pair (bp) perfect inverted repeat sequence at its ends. The terminal five base pairs (5' TGTTG . . . CAACA 3') are identical to the termini of Drosophila copia transposable elements. Two other Cin1 alleles are flanked by 5-bp direct repeats. A comparison is made with the long terminal repeat (LTR) of the copia-Ty1-retrovirus families of moveable genetic elements.     

Transposition of hAT elements links transposable elements and V(D)J recombination

Transposons are DNA sequences that encode functions that promote their movement to new locations in the genome. If unregulated, such movement could potentially insert additional DNA into genes, thereby disrupting gene expression and compromising an organism's viability. Transposable elements are classified by their transposition mechanisms and by the transposases that mediate their movement. The mechanism of movement of the eukaryotic hAT superfamily elements was previously unknown, but the divergent sequence of hAT transposases from other elements suggested that these elements might use a distinct mechanism. Here we have analysed transposition of the insect hAT element Hermes in vitro. Like other transposons, Hermes excises from DNA via double-strand breaks between the donor-site DNA and the transposon ends, and the newly exposed transposon ends join to the target DNA. Interestingly, the ends of the donor double-strand breaks form hairpin intermediates, as observed during V(D)J recombination, the process which underlies the combinatorial formation of antigen receptor genes. Significant similarities exist in the catalytic amino acids of Hermes transposase, the V(D)J recombinase RAG, and retroviral integrase superfamily transposases, thereby linking the movement of transposable elements and V(D)J recombination.     

Retrovirus-induced de novo methylation of flanking host sequences correlates with gene inactivity

The pattern of DNA methylation changes during development of eukaryotes, and hypomethylation frequently correlates with gene expression (for reviews see refs 1-4). A causal relationship between hypermethylation and gene inactivity has been established for retroviral genomes which are methylated de novo when inserted into the germ line of mice (ref. 5; for review, see ref. 6). The mutual interaction of the provirus with the host genome can influence virus expression and can result in inactivation of the host gene by insertional mutagenesis. We report here that the insertion of a provirus can change the methylation pattern of the host DNA. Sequences flanking the provirus become methylated de novo within 1 kilobase (kb) of the integration site. In Mov-13 mice, which carry a lethal mutation of the alpha 1(I) collagen gene, de novo methylation of host DNA is associated with a change in chromatin conformation. This suggests that virus-induced DNA methylation can alter DNA-protein interactions and thereby interfere with correct gene activation during embryonic development.     

A miniature insertion element transposable in Microcystis sp. FACHB 854

A 186-bp sequence with imperfect terminal inverted repeats and target direct repeats but without any transposase-encoding capacity was found to be transposable in an isolate derived from Microcystis sp. FACHB 854. This miniature insertion element, designated as ISM854-1, and with its homologues present at least 10 copies in the genome of Microcystis FACHB 854, is inserted into the 8-bp long and AT-rich target sequences, but none or few in other Microcystis strains. A variant of ISM854-1, denoted ISM854-1A, has perfect inverted repeat sequences and may transpose in pairs in a structure like a composite transposon. This is the first report of non-autonomous transposition of a mini-IS in a cyanobacterium.     

Polymorphism near the rat prolactin gene caused by insertion of an Alu-like element

Primate Alu and rodent Alu-like elements comprise major families of mammalian small dispersed repetitive DNAs. These elements are repeated more than 10(5) times per haploid genome and are found between known genes, in introns and in satellite DNA. Their dispersion throughout the genome and the presence of directly repeated DNA sequences flanking the elements suggest, but do not prove, that they are capable of transposition. We describe here an allelic variation in the 5'-flanking region of the rat prolactin gene that offers the opportunity to examine the sequences of matching regions of two homologous chromosomes which differ in the presence of an Alu-like repetitive DNA element. Our findings support the hypothesis that these elements are integrated into the genome by generating short direct repeats of host DNA.     

Transposition of Tn554 does not generate a target duplication

Transposable elements from prokaryotic and eukaryotic organisms are discrete DNA segments bounded by inverted or directly repeated sequences that insert into non-homologous DNA in a reaction that is independent of the general recombination functions of the host. The mechanisms proposed generally involve a staggered double-stranded scission of the target DNA, ligation to the nicked ends of the transposable element, and replication of the element, resulting in the generation of a directly repeated oligonucleotide target sequence flanking the new copy of the element. Most transposons have a relatively low degree of target site specificity coupled with a low insertion frequency. Tn554, a Staphylococcus aureus transposon which specifies resistances to erythromycin and spectinomycin, displays an unusually high degree of insertion specificity. Tn554 transposes with high efficiency to a unique ('primary') site in the S. aureus chromosome and only rarely (less than 10(-6) per transductant) to other, secondary sites. We report here the nucleotide sequences surrounding the junctions of Tn554 in three independent 'primary' insertions and two 'secondary' insertions of the transposon. Two unusual features are revealed: first, the termini of Tn554 contain neither inverted nor directly repeated sequences. Second, transposition of Tn554 does not generate the short direct repeats of the target DNA that are characteristic of other transposable elements. These results suggest that the mechanism of Tn554 insertion may be significantly different from that of other transposons.     

Clusters of CpG dinucleotides implicated by nuclease hypersensitivity as control elements of housekeeping genes

DNA sequences of the X-chromosome-linked hypoxanthine phosphoribosyltransferase (HPRT) and glucose 6-phosphate dehydrogenase (G6PD) genes have revealed the presence of clusters of CpG dinucleotides, raising the possibility that such clusters are involved in the control of expression of these genes, which are expressed in all tissues. Although CpG clusters are not exclusive features of the X chromosome, the analysis of X-linked genes provides the means to determine whether CpG clusters are control elements; one of the two homologous X loci in female mammals is not expressed, so that active and inactive versions of the gene can be compared. In fact, it has been shown that these CpG clusters are undermethylated when the gene is active and extensively methylated when the gene is inactive. In addition to hypomethylation, chromatin hypersensitivity to endonuclease digestion is a known hallmark of regulatory sequences in eukaryotic genes. We report here that the CpG clusters of the active hprt and g6pd genes are not only undermethylated, but also hypersensitive to MspI, DNase I and S1 nuclease, further supporting the suggestion that they are involved in the control of expression of these genes.     

Mobilization of a Drosophila transposon in the Caenorhabditis elegans germ line.

Transposons have been enormously useful for genetic analysis in both Drosophila and bacteria. Mutagenic insertions constitute molecular tags that are used to rapidly clone the mutated gene. Such techniques would be especially advantageous in the nematode Caenorhabditis elegans, as the entire sequence of the genome has been determined. Several different types of endogenous transposons are present in C. elegans, and these can be mobilized in mutator strains (reviewed in ref. 1). Unfortunately, use of these native transposons for regulated transposition in C. elegans is limited. First, all strains contain multiple copies of these transposons and thus new insertions do not provide unique tags. Second, mutator strains tend to activate the transposition of several classes of transposons, so that the type of transposon associated with a particular mutation is not known. Here we demonstrate that the Drosophila mariner element Mos1 can be mobilized in C. elegans. First, efficient mobilization of Mos1 is possible in somatic cells. Second, heritable insertions of the transposon can be generated in the germ line. Third, genes that have been mutated by insertion can be rapidly identified using inverse polymerase chain reaction. Fourth, these insertions can subsequently be remobilized to generate deletion and frameshift mutations by imperfect excision.     

A unique regulatory phase of DNA methylation in the early mammalian embryo

DNA methylation is highly dynamic during mammalian embryogenesis. It is broadly accepted that the paternal genome is actively depleted of 5-methylcytosine at fertilization, followed by passive loss that reaches a minimum at the blastocyst stage. However, this model is based on limited data, and so far no base-resolution maps exist to support and refine it. Here we generate genome-scale DNA methylation maps in mouse gametes and from the zygote through post-implantation. We find that the oocyte already exhibits global hypomethylation, particularly at specific families of long interspersed element 1 and long terminal repeat retroelements, which are disparately methylated between gametes and have lower methylation values in the zygote than in sperm. Surprisingly, the oocyte contributes a unique set of differentially methylated regions (DMRs)--including many CpG island promoters--that are maintained in the early embryo but are lost upon specification and absent from somatic cells. In contrast, sperm-contributed DMRs are largely intergenic and become hypermethylated after the blastocyst stage. Our data provide a genome-scale, base-resolution timeline of DNA methylation in the pre-specified embryo, when this epigenetic modification is most dynamic, before returning to the canonical somatic pattern.     

Genetic effects of injection of Rous sarcoma virus DNA into polar plasm of early Drosophila melanogaster embryos

Retroviral proviruses and the transposable elements of eukaryotic genomes are structurally similar. The biological significance of eukaryotic transposable elements has not been examined extensively but it is known that, like prokaryotic transposons, these elements can induce mutations in adjacent genes and cause their transposition. It is of interest to determine whether retroviral proviruses have the same mutagenic and gene transposing ability as transposable elements, particularly because the retrovirus genome is assumed to have originated from transposable elements of lower eukaryotes. The transfer of DNA sequences into animal zygotes or embryos by microinjection is a promising experimental approach for eluxidating their functions: when foreign DNAs were introduced into a mouse germ line, mutations were induced and at least in some mice, the mutation was caused by the insertion of a retroviral sequence. We have introduced Rous sarcoma virus (RSV) DNA into a germ line of Drosophila melanogaster, and describe here the resultant genetic effects.     

Non-function of a Moloney murine leukaemia virus regulatory sequence in F9 embryonal carcinoma cells

Moloney murine leukaemia virus (M-MuLV) infection of embryonal carcinoma (EC) cells results in the integration of proviral DNA into the host cell genome, but not in virus production. One suggested explanation for the lack of viral gene expression in EC cells has been methylation of the integrated viral DNA. However, subsequent reports indicated that integration of the M-MuLV DNA occurs soon after infection, but that viral DNA methylation occurs considerably later. Nevertheless, viral gene expression is not observed even at early times. One possible explanation is that certain M-MuLV regulatory sequences do not function in EC cells. We now present evidence which supports this hypothesis.