Portal de Periódicos da CAPES

Yeast Retrotransposons: Finding a Nice Quiet Neighborhood

1998; Cell Press; Volume: 93; Issue: 7 Linguagem: Inglês

10.1016/s0092-8674(00)81450-6

1097-4172

Jef D. Boeke, Scott E. Devine,

Chromosomal and Genetic Variations

Retrotransposons are ancient and ubiquitous genome parasites. The habitat of these elements includes the chromosomes of most if not all eukaryotes. As such, these elements, and particularly those that inhabit streamlined genomes such as those of yeasts, must exercise great caution in inserting into new sites. Those that fail to do so doom themselves and their progeny to extinction. Unlike infectious retroviruses, which can desert their individual host organisms via horizontal transmission, retrotransposons are permanently linked to the genome of their host and its vertically derived progeny (3Craigie R. Trends Genet. 1992; 8: 187-189Abstract Full Text PDF PubMed Scopus (89) Google Scholarreferences therein). Examination of the genome structures of yeast, flies, mammals, and plants reveals a striking correlation between genome size, coding information content, and transposon abundance (Figure 1). Simpler genomes such as that of Saccharomyces cerevisiae are densely packed with ORFs and contain relatively little transposon sequence. As genome complexity increases, the percent of the genome committed to encoding cellular proteins declines and the abundance of transposons, and especially retrotransposons, increases. Presumably, jumping into a repetitive sequence, such as another transposon copy, usually has little or no impact on the survival of the host organism. The incremental metabolic load conferred by adding a few kilobases to an already enormous host genome is negligible. Thus, on a random basis, the probability of long-term survival (i.e., not hitting a gene important for host survival) is much higher for those retrotransposons that inhabit the larger genomes of multicellular organisms than they are for those of the minimalist yeast genomes. It follows that the strongest selection for evolution of mechanisms enabling host gene avoidance will have been imposed in organisms with smaller genomes like yeasts. An interesting corollary to this hypothesis is that the relatively sloppy elements that lack the precise integration mechanism afforded by LTRs (long terminal repeats) appear to have been eliminated from S. cerevisiae's streamlined genome. The latter elements, typified by the mammalian L1 and Alu sequences, can reach enormous abundance in mammalian genomes, where they are estimated to constitute as much as 30% of the total DNA by weight (15Smit A.F.A. Curr. Opin. Genet. Devel. 1996; 6: 743-748Crossref PubMed Scopus (474) Google Scholar). It is estimated that we humans carry around several ounces of retrotransposon chromatin, approximately equivalent in weight to a Big Mac—with fries. The great bulk of the L1 element copies are believed to be inactive due to deletions, inversions, and point mutations in the sequences. In the few cases in which these "sloppier" LTR-less retroelements have been found in unicellular eukaryotes, they are of relatively low abundance. There are now several spectacular examples of how yeast retrotransposons, or Ty, elements avert catastrophic insertion events. All of these Ty elements are of the LTR type. There appear to be at least three distinct types of strategies they use to dodge host genes; these have been studied in the Ty1, Ty3, and Ty5 elements (Table 1). Ty1 and Ty3 utilize apparently different mechanisms to target genomic regions lying immediately upstream of genes transcribed by RNA polymerase III, and Ty5 targets transcriptionally "silent" regions of the yeast genome. Both of these strategies home in on nonessential gene-poor regions of the yeast genome.Table 1Targeting by Ty Elements in the Yeast GenomeTy element typePhylogenetic ClassificationFavored Target(s)/ Proposed RecognitionTy1Ty1-copiapol III genes/chromatinaDevine and Boeke 1996.Ty3Ty3-gypsypol III genes/transcription factorbKirchner et al. 1995.Ty5Ty1-copiasilent loci, telomeres/chromatincZou et al. 1996.a 5Devine S.E. Boeke J.D. Genes Dev. 1996; 10: 620-633Crossref PubMed Scopus (172) Google Scholar.b 10Kirchner J. Connolly C.M. Sandmeyer S.B. Science. 1995; 267: 1488-1491Crossref PubMed Scopus (178) Google Scholar.c 19Zou S. Ke N. Kim J.M. Voytas D.F. Genes Dev. 1996; 10: 634-645Crossref PubMed Scopus (153) Google Scholar. Open table in a new tab All three types of Ty elements share an overall life cycle that in outline resembles that of the retroviruses, except that there is no extracellular or infectious phase. Like retroviruses, the termini of Ty elements are defined by LTR sequences. These serve to define both the Ty RNA 5′ and 3′ ends as well as the extreme Ty DNA termini that are directly involved in the integration reaction that joins Ty and host genome sequences. Ty elements encode gag and pol genes, and as in retroviruses, pol encodes protease, reverse transcriptase, and integrase functions, all of which are essential for retrotransposition. Ty RNA is translated and also packaged into the virus-like particles, or VLPs, which are direct retrotransposition intermediates. Specific cleavages in the Gag and Gag-Pol proteins made by the Ty protease produce the mature capsid, protease, reverse transcriptase, and integrase proteins. Reverse transcription of Ty RNA is carried out in the VLPs, and the cDNA produced is then transported to the host cell nucleus. The transported entity probably consists of a Ty cDNA/integrase complex; Ty1 integrase was recently shown to contain a nuclear targeting signal (8Kenna M.A. Brachmann C.B. Devine S.E. Boeke J.D. Mol. Cell. Biol. 1998; 18: 1115-1124PubMed Google Scholar, 13Moore S.P. Rinckel L.A. Garfinkel D.J. Mol. Cell. Biol. 1998; 18: 1105-1114PubMed Google Scholar). However, the overall biochemical identity of this preintegration complex (PIC) is still unknown; its composition should provide important clues to deciphering the mechanism by which genomic integration sites are chosen. The Ty3 element targets tRNA genes and other pol III–transcribed genes with startling precision; nearly all Ty3 elements in the yeast genome, as well as new Ty3 hops generated experimentally, are inserted within a few base pairs of the pol III transcription start site (2Chalker D.L. Sandmeyer S.B. Genes Dev. 1992; 6: 117-128Crossref PubMed Scopus (156) Google Scholar). As these start sites do not have a conserved sequence, and insertion requires pol III transcription factors, it is likely that this targeting is mediated by direct interactions between the Ty3 PIC and one of these factors (10Kirchner J. Connolly C.M. Sandmeyer S.B. Science. 1995; 267: 1488-1491Crossref PubMed Scopus (178) Google Scholar). Ty1 also targets pol III–transcribed genes, but the pattern of insertions is very different. An "integration window" for Ty1 insertion, extending from about 75–700 bp upstream of the transcription start site is seen for diverse pol III gene targets. As with Ty3, pol III gene targeting by Ty1 requires an intact pol III promoter at the target site, implicating pol III transcription in the targeting mechanism. Nevertheless, the shape of the integration window suggests that integration occurs within the context of nucleosomes normally present upstream of these pol III targets (5Devine S.E. Boeke J.D. Genes Dev. 1996; 10: 620-633Crossref PubMed Scopus (172) Google Scholar). The pol III–transcribed gene upstream region corresponding to the integration window is a domain that has been associated with down-regulation of pol II promoters intentionally placed there (7Hull M. Erickson Johnston M. Engelke D. Mol. Cell Biol. 1994; 14: 1266-1277Crossref PubMed Scopus (98) Google Scholar), possibly due to the effects of pol III gene nucleosome positioning signals (14Morse R.H. Roth S.Y. Simpson R.T. Mol. Cell. Biol. 1992; 12: 4015-4025Crossref PubMed Scopus (69) Google Scholar, 4Curcio M.J. Morse R.H. Trends Genet. 1996; 12: 436-438Abstract Full Text PDF PubMed Scopus (29) Google Scholar). Studies from Sentenac's lab suggest that nucleosomes are positioned in the corresponding U6 gene region (U6 is transcribed by pol III in yeast) (12Marsolier M.C. Tanaka S. Livingstone-Zatchej M. Grunstein M. Thoma F. Sentenac A. Genes Dev. 1996; 15: 410-422Google Scholar). Ty1 is known to insert quite readily into the classically silenced HM loci (18Weinstock K. Mastrangelo M. Burkett T. Garfinkel D. Strathern J. Mol. Cell. Biol. 1990; 10: 2882-2892PubMed Google Scholar) and into the "neoclassically" silenced rDNA (1Bryk M. Banerjee M. Murphy M. Knudsen K.E. Garfinkel D.J. Curcio M.J. Genes Dev. 1997; 11: 255-269Crossref PubMed Scopus (324) Google Scholar, 16Smith J.S. Boeke J.D. Genes Dev. 1997; 11: 241-254Crossref PubMed Scopus (495) Google Scholar) as well as upstream of pol III genes. These and other results suggest that the Ty1 preintegration complex may interact with a chromatin component(s) to specify targeting. The apparently distinct mechanisms (inferred from the distinct insertion patterns and other data) by which Ty1 and Ty3 target pol III upstream regions suggest convergent evolution of targeting for these genomic regions. In the yeast genome, regions upstream of pol III–transcribed genes are typically gene-free, presumably because of the deleterious effects of the pol III–transcribed gene on pol II transcription (7Hull M. Erickson Johnston M. Engelke D. Mol. Cell Biol. 1994; 14: 1266-1277Crossref PubMed Scopus (98) Google Scholar), and thus represent "safe havens" into which a Ty can insert without compromising the health of its host (Figure 2). The Ty5 element on the other hand shows a clear-cut insertion preference for classical silent chromatin; about 95% of natural Ty5 insertions and experimentally induced ones were found either at telomeres or at the silent HM loci (19Zou S. Ke N. Kim J.M. Voytas D.F. Genes Dev. 1996; 10: 634-645Crossref PubMed Scopus (153) Google Scholar). As these classical silent loci are definitively associated with altered chromatin structure, some aspect of that structural alteration may represent a biochemical signal recognized by Ty5. As Ty1 and Ty5 are much more closely related to each other than to Ty3, it makes sense that they might share some fundamental mechanism like the biochemistry of target site definition. The general problem of biological specificity can be elegantly tackled by the isolation of loss of specificity mutants, which provide an entree into the recognition mechanism. In the current issue of Molecular Cell, 6Gai X. Voytas D.F. Mol. Cell. 1998; 1: 1051-1055Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar report just that. A new assay for Ty5-HIS3 was developed in which transposition into a specific preferred silent target, placed on an episome containing the (nonsilenced) color marker ADE2, gave rise to pure white His+ colonies, while transpositions elsewhere gave rise to red/sectored colonies. Following mutagenesis of a Ty5 donor plasmid, a single missense mutant was obtained in the pol region that gave rise to an ∼20-fold decrease in white colonies compared to the wild-type Ty5 donor. Analysis of 10 insertions generated by the mutant indicated that they had transposed to yeast chromosomes and none were in classical silent chromatin. In contrast, about 95% of previously studied experimentally induced Ty5 hops were found in silent regions. Thus, this mutation clearly causes Ty5 to lose specificity for silent DNA. The missense mutation is in the pol region, which encodes at least one central protein component of the PIC, integrase itself. However, the data provided do not allow an unambiguous conclusion to be reached about which Ty5 protein(s) are affected by the mutation, which lies perilously close to the integrase/reverse transcriptase boundary. A multiple sequence alignment of several related yeast transposons supports the idea that the mutation maps to the extreme C terminus of integrase. However, due to the rather limited interelement homology within this coding region (rather extensive gapping was required), the alignment presented may not be optimal and the mutation might end up affecting the N terminus of reverse transcriptase or possibly even the protease cleavage site separating the two proteins. It is not unreasonable that reverse transcriptase might play a role in late steps of Ty transposition (e.g., in terminal gap repair), and therefore it is possible that reverse transcriptase could be part of the PIC. Biochemical identification of the N terminus of reverse transcriptase is needed to resolve the issue of what Ty5 protein is affected by the missense mutation. The mutation presumably defines a part of a Ty5 Pol protein that interacts with a host component important for targeting integration. An obvious possibility is that the altered component is integrase, and that this part of integrase interacts directly or indirectly with some (presumably host-encoded) component of silent chromatin. If so, it may be possible to identify that component by selecting for suppressors of the Ty5 missense mutation in host genes. If the presented alignment is correct, it places the Ty5 mutation near the position of a recently identified Ty1 nuclear localization signal in integrase (8Kenna M.A. Brachmann C.B. Devine S.E. Boeke J.D. Mol. Cell. Biol. 1998; 18: 1115-1124PubMed Google Scholar, 13Moore S.P. Rinckel L.A. Garfinkel D.J. Mol. Cell. Biol. 1998; 18: 1105-1114PubMed Google Scholar). Because the yeast nuclear membrane remains intact throughout the cell cycle, Ty elements, like HIV virions infecting nondividing cells, must encode special machinery to traverse this intracellular barrier. If the Ty5 mutation has actually hit a nuclear targeting signal, this raises the intriguing possibility that there may be different classes of nuclear targeting signals, perhaps targeting different intranuclear compartments. The idea that silent chromatin occupies a separate perinuclear "compartment" is supported by a variety of genetic and cell biological studies (17Wakimoto B. Cell. 1998; 93: 321-324Abstract Full Text Full Text PDF PubMed Scopus (183) Google Scholarreferences therein). Such a silent chromatin compartment(s) may not meet a strict cell biological definition, i.e., a compartment fully enclosed by a lipid bilayer. Nevertheless, the bulk of telomeric DNA, rDNA, and classical Drosophila heterochromatic sequences seems to be isolated physically from the rest of the DNA sequences and bound on one side by nuclear membrane. Might there be specialized portals of entry to this compartment that are being exploited by Ty5 in order to direct integration preferentially into silent chromatin? If this were the mechanism, then a loss of the ability to target this compartment would either reduce transposition frequency by 95% (if Ty5 could only get into the nucleus efficiently by this route) or it could go in at 100% efficiency if there were an efficient backup pathway. The mutant is about 30% as efficient as the wild type, suggesting the latter, perhaps in combination with a mild deleterious pleiotropic effect on transposition. The isolation of this mutant lays a firm foundation for the notion that Tys play an active role in finding their targets; in the Ty5 case, the silent chromatin. Regardless of the specific mechanism of the targeting defect in this particular missense mutant, Ty elements are likely to offer new tools and new insights to help "demystify" silencing mechanisms (11Lewin B. Cell. 1998; 93: 301-303Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar).