Transposable elements (TEs) are abundant in the genomes of eukaryotes, making up >40% of the nuclear DNA of humans (1). Many TEs appear to be simply selfish DNA “whose primary and often only function is self-preservation” (2). Of course, selfish behavior has effects on bystanders, and TEs can have a number of effects on the rest of the genome. TEs can affect gene expression by inserting within a gene and disrupting its structure, or by directing transcription of a nearby gene. TEs also can affect splicing and transcription termination, thereby altering the protein product of a gene. TEs can insert at multiple sites in the genome, allowing recombination between otherwise nonhomologous regions, resulting in large-scale rearrangements of chromosomes. In fact, many of the best-known TEs were initially identified because they mutated a gene under study. TEs can even have incidental benefits for the host, and there is one example of TEs devoted entirely to the benefit of the genome in which they reside, the Drosophila telomeric transposons, HeT-A and TART. In this issue of PNAS, Arkhipova and Morrison (3) report two very interesting TEs from Giardia lamblia that may also have evolved to benefit their host. This study adds another piece to the emerging picture of what happens at chromosome ends.
Although genomes have evolved measures to keep TEs in check (4), Hickey (5) has pointed out that even deleterious TEs can spread through sexual populations because of their ability to reproduce and colonize both homologous chromosomes in the zygote. Thus, the reduction in reproductive fitness of the host can be balanced by the transmission of TEs to all its progeny. In contrast, in asexual populations, elements depend entirely on survival of their host, suggesting that deleterious TEs are unlikely to persist in asexual organisms. Arkhipova and Meselson (6) provided support for this idea in a study showing that the Bdelloidea, a class of rotifers that apparently became asexual millions of years ago, completely lacks retrotransposons.
TEs in Giardia
Arkhipova and Morrison (3) examine the three TE families GilM, GilT, and GilD found in another putatively asexual species, the parasitic protozoan, G. lamblia. All three are non-long terminal repeat (non-LTR) retrotransposons of the class typified by the human LINE-1 element. They conclude that the G. lamblia elements are either dead or safely sequestered, supporting the suggestion that asexual species cannot maintain deleterious TEs. The GilD family consists of highly diverged copies that appear to be relics of dead elements. The other two families, GilM and GilT, share unusual properties: they are found only in subtelomeric tandem head-to-tail arrays between the single copy genes and terminal arrays of (TAGGG)n added by telomerase. As the authors point out, this subterminal location diminishes the possibility that GilM and GilT will interfere with gene function and may confer a benefit by expanding the buffer between the genes and the end of the chromosome. It is interesting to consider the possibility that these elements have coevolved with the host to assume other roles in the chromosome as well.
In response to environmental pressures, the chromosomes of Giardia demonstrate remarkable length plasticity.
G. lamblia is a protozoan with two polyploid nuclei that appear to be functionally equivalent (7). In response to environmental pressures, the chromosomes of Giardia demonstrate remarkable length plasticity. For example, chromosome 1 can vary from 1.1 to 1.9 Mb. The plasticity is confined primarily to subtelomeric regions; the central regions are much less variable. Some of the change is caused by changes in ribosomal genes (rDNA); however, the TEs are a major component of the subtelomeric region and also may be involved in amplification and loss.
The plasticity of the G. lamblia genome, the marked changes in subtelomeric regions, and the response to environmental factors suggest analogy to the ciliated protozoa, which also have two nuclei (8, 9). Ciliates' have one nucleus, the micronucleus, which contains the complete genome and undergoes conventional mitoses and meioses. It apparently constitutes the germ line of the organism. The other nucleus, the macrocronucleus, is rebuilt from the micronucleus after each mating. Although details differ in species-specific ways, this rebuilding involves an elaborate program that discards many sequences, amplifies others, and adds telomeres to the new ends created by chromosome fragmentation. The resulting macronucleus is highly polyploid and carries the cell through many generations of vegetative growth—until the next sexual cycle, an event that is responsive to the environment. Our understanding of the biochemistry of telomeres, including the discovery of telomerase (10), began with the study of ciliated protozoa.
Are there analogies between the two nuclei of G. lamblia and the two nuclei in ciliates? For instance, are the G. lamblia TEs subtelomeric because the cells have a mechanism analogous to that used in constructing the ciliate macronucleus, which facilitates the elimination of sequences and readdition of TAGGG repeats? Such a mechanism could allow the TEs to be amplified and deleted rapidly in response to environmental changes. This speculation suggests a number of experiments. To what extent are the G. lamblia TE sequences involved in the subtelomeric changes in chromosome size? Do these TE sequences have special roles when amplified, as some of the ciliate micronuclear-specific sequences have in meiosis? It would be interesting to know how environmental pressures affect the number of TE sequences and what effect variations in the number of TE sequences have on the cell.
The analogy suggested above is not entirely speculative. In Saccharomyces cerevisiae, multiple copies of genes adaptive for certain environments are associated with chromosome ends, suggesting that this location facilitates their amplification (11). The functional importance of these genes is easy to see because we understand the gene products. The roles of the TE sequences in Giardia are more cryptic but they, too, might be exploiting the ease with which gene amplification and diminution can occur at chromosome ends.
Giardia and Drosophila
Although GilM and GilT have not replaced telomerase telomeres, it is possible, even likely, that Giardia and its subtelomeric TEs have coevolved a mutually beneficial symbiosis (see below). In contrast, Drosophila uses retrotransposons, HeT-A and TART, as telomeres (12, 13). Telomeres are proving to have an increasing number of functions, many of which we do not understand (12, 14). Although the Giardia elements clearly do not fulfill all telomeric functions, they may share some functions with the Drosophila telomeric elements. Most techniques of genetic and reverse genetic analysis are powerless to test questions about multicopy elements like telomere repeats, but comparisons of species with variant telomeres offer an opportunity to extend our understanding. It is thus especially interesting to carefully compare the G. lamblia elements GilM and GilT with the Drosophila telomere transposons HeT-A and TART.
HeT-A, TART, GilM, and GilT are all non-LTR retrotransposons, a class that transposes by a poly(A) RNA intermediate that is reverse-transcribed directly onto chromosomal DNA (15). When non-LTR elements transpose into internal sites, reverse transcription is primed off a nick in the DNA. We believe that HeT-A and TART are primed off the extreme end of the chromosome. A comparison of the structures of the four TEs (Fig. 1 and the sequence analysis in ref. 3) shows that GilM and GilT are closely related and very distantly related to the Drosophila TEs. In contrast, HeT-A and TART have different origins and appear to be products of convergent evolution (16). We have suggested that HeT-A and TART are targeted to the telomeres because they have acquired related Gag proteins. These proteins show no obvious similarity to the nucleic acid binding proteins of the G. lamblia TEs, suggesting that Gil's subtelomeric targeting involves a different mechanism.
Figure 1.
Structure of non-LTR retrotransposons. Members of the Giardia subtelomeric TE families, GilM and GilT, and the Drosophila telomeric TEs, HeT-A and TART are diagrammed. LINE-1 of humans is shown for comparison. ORFs are color-coded by function: purple, nucleic acid binding; turquoise, reverse transcriptase; green, endonuclease. Zinc knuckle (CCHC) and zinc finger (CCHH) motifs are in red and salmon, respectively. Only LINE-1 has identifiable target site duplications (red triangles). The irregular pattern of repeats in the 3′ UTR of HeT-A is indicated by shading. Black bars below TART indicate a distinctive pair of perfect nonterminal repeats (PNTR) that characterize a subfamily of non-LTR elements. TART is the only TE in this figure with such repeats. Pink arrows indicate variable lengths of oligo(A) found at the 3′ end of every copy of these elements.
A striking similarity between the structures of the G. lamblia and Drosophila TEs is the very long 3′ untranslated region (UTR). Almost all TEs devote most of their sequence to coding for proteins necessary for their own transposition. It seems reasonable that TEs, like viruses, would find it advantageous to minimize the length of sequence that needs to be moved during transposition. Thus it is surprising to see that GilM, GilT, HeT-A, and TART all have 3′ noncoding sequences that are nearly as long as the coding region (Fig. 1). This feature suggests convergent evolution toward some shared functionality. The 3′ UTR of HeT-A has an irregular pattern of A-rich repeats throughout. This pattern, but not the specific sequence, is conserved in the HeT-A elements of related Drosophila species (17). It is likely that this conserved pattern reflects a function, possibly binding specific proteins to establish the heterochromatic chromatin in telomere regions. The 3′ UTRs of TART, GilM, and GilT lack this dramatic sequence pattern but still may have important roles in protein binding.
The second striking similarity between the G. lamblia TEs and the Drosophila telomere transposons is in their ability to form head-to-tail arrays of specific polarity at the chromosome end. However, the arrays of Giardia and Drosophila have several obvious differences, which may reflect different transposition mechanisms. In contrast to HeT-A and TART, GilM and GilT are not on the extreme end and are never found in mixed arrays, internal elements in the arrays are always complete, and the most distal element is invariably truncated by the TAGGG cap. Both the Gil arrays also lack target site duplications between elements, as do the Drosophila arrays.
The Drosophila elements form the extreme ends of the chromosomes. G. lamblia clearly has a robust telomerase and all Giardia TE arrays reported are capped by long arrays of telomere repeats. Neither are these Giardia TEs essential for all chromosomes; telomere repeats are joined to rDNA or surface protein (VSP) genes in the other three sequences in table 1 of ref. 3. Any explanation for the production of these arrays must take the presence of TAGGG caps into consideration.
For HeT-A and TART, studies of both normal and healed broken telomeres argue strongly that the polar arrays form by repeated reverse transcription primed by the 3′ OH on the end of the chromosome. This mechanism forces the same orientation on all members of the array. (It is important to note that this mechanism is not target sequence-specific but does appear to be specific for HeT-A and TART, which form long mixed arrays. None of the many other known Drosophila non-LTR elements are ever found in the telomere, nor are HeT-A or TART ever found outside it.)
Arkhipova and Morrison (3) offer two possible mechanisms for generating the polar arrays of G. lamblia TEs. Both require interactions between telomerase and the TEs. We suggest another version can explain the known features more economically. We propose that the invariant polarity is caused by sequence-specific insertion of TEs. (Other subtelomeric elements producing fixed polarity by such insertion are discussed below.) As in one of the explanations suggested in ref. 3, we propose that the target is the 3′ UTR of the element itself. Sequence-specific transposition into its own 3′ UTR could provide polarity and explain why each array contains a single family. Insertion in the poly(A) region would make target site duplication indistinguishable from the poly(A) of the next element. As in macronucleus building in ciliates, chromosomal shortening could result from programmed breakage followed by capping with telomerase. Truncation of the distal TE and shortening of the subtelomeric array thus would be the result of a chromosome break. Expansion of the array could occur by insertions into the 3′ UTR of any element, which would leave a broken element at the distal end of the array, under the TAGGG cap. The three junctions shown in table 1 of ref. 3, in which telomerase has capped ribosomal genes or genes for surface proteins, could result from either more vigorous end erosion or size reduction of chromosomes without sub terminal TEs, followed by telomerase capping.
Other Species
Three non-LTR elements known to transpose to telomere regions (boxed in figure 3 of ref. 3) show that specific targeting can generate insertions of specific polarity in subtelomeric regions (13). TRAS1 and SART1, from the silkworm, are targeted to specific nucleotides in the TTAGG repeats and insert in opposite orientations. The position and direction of insertion are determined by the transposition machinery rather than any nucleic acid complementarity between TE sequence and target site. Importantly, the general location of the insertions must be under additional control, perhaps from chromatin structure, because insertions do not occur within 6–8 kbp of the extreme end of the chromosome despite the abundance of DNA targets in these regions. The third element, Zepp, from Chlorella, integrates into itself, producing head-to-tail arrays, some of which have been found near telomeres.
Conclusion
Our understanding of human telomeres and telomerase has benefited from studies of telomeres from ciliated protozoa and yeast. The method by which G. lamblia maintains its subtelomeric TE arrays must be significantly different from the method by which Drosophila maintains its telomeric arrays. Nevertheless, these arrays may have some analogous functions. Information gleaned from each will be important to understanding both telomeres and TEs.
Footnotes
See companion article on page 14497.
References
- 1.Smit A F. Curr Opin Genet Dev. 1999;9:657–663. doi: 10.1016/s0959-437x(99)00031-3. [DOI] [PubMed] [Google Scholar]
- 2.Doolittle W F, Sapienza C. Nature (London) 1980;284:601–603. doi: 10.1038/284601a0. [DOI] [PubMed] [Google Scholar]
- 3.Arkhipova I R, Morrison H G. Proc Natl Acad Sci USA. 2001;98:14497–14502. doi: 10.1073/pnas.231494798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Waterhouse P M, Wang M B, Lough T. Nature (London) 2001;411:834–842. doi: 10.1038/35081168. [DOI] [PubMed] [Google Scholar]
- 5.Hickey D A. Genetics. 1982;101:519–531. doi: 10.1093/genetics/101.3-4.519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Arkhipova I, Meselson M. Proc Natl Acad Sci USA. 2000;97:14473–14477. doi: 10.1073/pnas.97.26.14473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Adams R. Clin Microbiol Rev. 2001;14:447–475. doi: 10.1128/CMR.14.3.447-475.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Gall J G, editor. Molecular Biology of Ciliated Protozoans. New York: Academic; 1986. [Google Scholar]
- 9.Katz L A. Int J Syst Evol Microbiol. 2001;51:1587–1592. doi: 10.1099/00207713-51-4-1587. [DOI] [PubMed] [Google Scholar]
- 10.Greider C W, Blackburn E H. Cell. 1985;43:405–413. doi: 10.1016/0092-8674(85)90170-9. [DOI] [PubMed] [Google Scholar]
- 11.Ness F, Aigle M. Genetics. 1995;140:945–956. doi: 10.1093/genetics/140.3.945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pardue M-L, DeBaryshe P G. Chromosoma. 1999;108:73–82. doi: 10.1007/s004120050354. [DOI] [PubMed] [Google Scholar]
- 13.Pardue M-L, DeBaryshe P G. In: Mobile DNA II. Craig N, Craigie R, Gellert M, Lambowitz A, editors. Washington, DC: Am. Soc. Microbiol.; 2002. , in press. [Google Scholar]
- 14.Blackburn E H. Cell. 2001;106:661–673. doi: 10.1016/s0092-8674(01)00492-5. [DOI] [PubMed] [Google Scholar]
- 15.Luan D D, Korman M H, Jakubczak J L, Eickbush T H. Cell. 1993;72:595–605. doi: 10.1016/0092-8674(93)90078-5. [DOI] [PubMed] [Google Scholar]
- 16.Danilevskaya O N, Traverse K L, Hogan N C, DeBaryshe P G, Pardue M-L. Mol Cell Biol. 1999;19:873–881. doi: 10.1128/mcb.19.1.873. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Danilevskaya O N, Tan C, Wong J, Alibhai M, Pardue M-L. Proc Natl Acad Sci USA. 1996;95:3770–3775. doi: 10.1073/pnas.95.7.3770. [DOI] [PMC free article] [PubMed] [Google Scholar]