Mary-Lou Pardue

Telomeres (the ends of chromosomes) have important roles in chromosome replication, in cell division, and in the cell-type-specific architecture of interphase nuclei. Telomere biology is fundamental to important aspects of cell biology, including cellular senescence, cell cycle checkpointing, organismal aging, and tumorigenesis, but we still know little about the mechanisms involved. Telomeres are complex, dynamic nucleoprotein structures formed on long arrays of repeated DNA sequences. In most organisms telomeres are maintained by an enzyme, telomerase. This enzyme compensates for erosion of chromosome ends by adding new repeats to the telomere array. These repeats are very short DNA sequences copied from the enzyme’s RNA template.

We study Drosophila telomeres, the original genetic and cytological model for telomeres. Thus we were surprised to discover that Drosophila telomeres are not typical telomerase-maintained structures. Instead they are maintained, not by telomerase, but by special transposable elements, the retrotransposons, HeT-A, TART, and TAHRE. Nevertheless, Drosophila telomeres are functionally similar to telomeres in other organisms. They are also structurally more similar to other telomeres than they might appear at first glance. The retrotransposons transpose by means of a poly(A)+ RNA which is reverse-transcribed directly onto the end of the chromosome. Successive transpositions form long arrays of head-to-tail repeats. The repeats are analogous to the repeats added by telomerase but Drosophila repeats are copies of the retrotransposons, three orders of magnitude longer than the repeats added by telomerase. Although individual repeats on Drosophila telomeres are very long, the repeat arrays on Drosophila chromosome ends are similar in length to telomere arrays in other multicellular organisms (many kilobases per end). Because both the retrotransposons and telomerase extend telomeres by adding copies of an RNA template, and because each type of telomere interacts functionally with homologues of many of the same proteins, the two mechanisms are basically variants of one another. Thus the variant Drosophila telomere provides an important model to investigate difficult questions of telomere biology.

The Drosophila telomeric retrotransposons are also unusual transposable elements. They are the first such elements shown to be entirely devoted to an essential cellular function. (Like their close relatives, the retroviruses, transposable elements are generally considered to be parasitic.) Drosophila telomeres provide a link between telomeres and transposable elements that raises interesting questions about the evolution of both eukaryotic chromosomes and transposable elements.


Model for extension of Drosophila telomeresOur model for the extension of Drosophila telomeres by the telomeric retrotransposons, HeT-A and TART.  Arrows represent the head-to-tail array of HeT-A (blue) and TART (green) on the chromosome end. (Actually each chromosome has many more copies of these elements; some copies are complete and some are 5’ truncated). Transcription of a member of this array provides a sense strand RNA (purple) that is translated in the cytoplasm to yield Gag protein (from either HeT-A or TART. The Gag protein associates with the sense strand RNA to deliver it to the chromosome end where the RNA becomes the template for reverse transcription onto the end of the chromosome, adding a new element.  Analogy with retroviruses suggests that reverse transcriptase is also included in the Gag:RNA complex: however there is no evidence on this point.


Phylogenetic studies of telomere-specific retrotransposons

HeT-A, TART, and TAHRE are clearly related to a clade of non-LTR retrotransposons that are abundant in the genome of Drosophila, yet the telomeric elements are never found in euchromatic, gene-rich regions which are littered with their non-telomeric relatives. Conversely, the non-telomeric relatives are not found in telomere arrays. The telomeric elements have other distinctive features not found in their relatives (see below). To investigate the importance of these features for telomere biology, we have isolated telomere-specific elements from distantly related Drosophila species. Our studies of these Drosophila species show that the telomere-specific retrotransposons have been coevolving with their host cells since at least the separation of the most distantly related Drosophila species (40-60 My), if not significantly longer. Telomeric features are conserved and apparently related to the roles of the elements in telomeres.

Potential roles for Gag proteins in retroelement targeting

HeT-A, TART, and TAHRE transpose only to chromosome ends, including ends of broken chromosomes. Drosophila is infested by many othertransposons that transpose to many sites, including gene-rich regions where they may cause mutations, yet these are never found in the telomere arrays. All non-LTR elements transpose by the same mechanism, target primed reverse transcription, yet there is no evidence that any specific DNA sequence defines the unique transposition site of HeT-A, TART, and TAHRE; therefore their chromosomal distribution must be directed by other mechanisms. We have found that the Gag protein encoded by the telomere-specific retrotransposons is a major factor in targeting their transposition to chromosome ends. We analyzed intracellular targeting of Gag proteins from six non-LTR retrotransposons, HeT-A, TART, TAHRE, and three non-telomeric elements, jockey, Doc, and I factor. These Gag proteins have high levels of sequence similarity, but they have dramatic differences in intracellular targeting. As expected, HeT-A and TART Gags were efficiently transported to nuclei where HeT-A Gag localized at telomeres and directed both TART and TAHRE Gags to telomeres. TART’s role in this cooperative effort may be supplying reverse transcriptase, an enzyme HeT-A lacks.

In contrast to the easy entry of telomeric retrotransposons into the nucleus, Gags of non-telomeric retrotransposons remained in the cytoplasm. These experiments demonstrate that closely related retrotransposon Gag proteins can have different intracellular localizations, presumably because they interact differently with cellular components. We suggest that these localizations reflect coevolution of the retrotransposons and their host cells to facilitate transposition of the telomere elements while hindering transposition of the parasitic elements.

Sequences of assembled telomere arrays record the history of events on the ends of chromosomes

Because telomere arrays are formed by successive transpositions of retrotransposons, each element is younger than its proximal neighbor and the distal 5’ end of each element is at risk of terminal erosion until another element transposes to protect it. Unfortunately, this historical information is difficult to obtain because the complex repeated sequences in telomere arrays preclude accurate assembly by whole genome sequencing so most available sequences are not useful. We were finally able to analyze this history when the Drosophila Heterochromatin Genome Project offered us an opportunity to collaborate on directed assembly of >70kb (kilobases) of telomere array from D. melanogaster chromosome 4R and >20kb from chromosome XL. Neither sequence extends to the chromosome terminus but each is linked to subtelomere sequence and therefore contains the oldest elements in its array. These sequences provided the first data for a solid quantitative description of D. melanogaster telomere arrays. Both arrays are composed entirely of HeT-A and TART elements, each with its 5’ end toward the end of the chromosome, as expected if they were reverse-transcribed onto the 3’OH on the end of chromosomal DNA. There were some surprises. Consistent with the role of telomeres in chromosome end protection, many of the retrotransposons are 5’-truncated, some severely; however, a surprisingly large fraction of elements are complete. Also unexpectedly, even the oldest elements deep in the array showed no evidence of the sequence decay one might expect of DNA that is no longer under selection for function.. Although these undecayed, intact, transposition-competent elements were unexpected, some such elements are clearly necessary as breeding stock for new elements to continue telomere maintenance.

The assembled telomere arrays provide sufficient data for quantitative analysis of events at the ends of telomere arrays

Several groups have shown that terminally deleted Drosophila chromosomes that completely lack telomeres shorten at a relatively regular rate of 50-100 nucleotides per fly generation, approximately the rate predicted by the well-known “end replication problem”. However these chromosomes lacking telomere sequences are rare. Many are eventually healed by transposition of new telomere transposons. The rest are lost when terminal erosion removes essential sequence. Our analyses allow us to conclude that, in contrast to the relatively regular loss on broken chromosomes, established telomeres undergo both small scale stochastic end erosion and more irregular large scale terminal deletions.

Our discovery that the telomeric retrotransposons have innovative ways to add expendable sequence to their 5’ ends (see below), led us to do a quantitative analysis of complete elements in telomere arrays that revealed small scale stochastic end erosion after each transposition. The added expendable sequence serves to buffer the essential sequence of the terminal element from this erosion. This buffering appears to be effective because complete elements are overrepresented in the telomere arrays. Moreover, all complete elements in those telomeres had been capped by a new transposition (that replaced them as the extreme end of the array) before erosion removed all of the non-essential sequence.

Small scale erosion is not the only cause of sequence loss from telomere arrays. The 5’ truncated partial elements in the arrays have a quantitative loss distribution which provides evidence of large scale terminal loss. Thus, our analyses allow us to postulate the existence of a complex process that maintains telomere length homeostasis while providing a supply of transposition-competent elements. The small scale erosion, seen at the 5’ end of complete elements removes very few nucleotides compared to those added by transposition of a new element (each new element is 6kb-13kb long). This presents a problem for maintaining telomere length homeostasis. The large scale terminal loss revealed by the 5’truncated elements provides a solution. We suggest that telomere length is restored by occasional terminal deletions that remove part, or sometimes all, of the telomere array, followed by rapid rebuilding of the array to the required length. Rapid rebuilding would help explain another unexpected finding of our study: elements deep in the array should have been there for a long time without selection for function but all appear to have undecayed sequence. Terminal deletions would remove decayed elements and rebuilding an array would of necessity add transposition-competent elements.

It is interesting that the number of proteins and activities that are still being recognized in human telomeres suggest that their maintenance may also be more complex than we now know. Perhaps the retrotransposon and telomerase variants will prove to have even more similarity than we now recognize.

Telomere-specific retrotransposons have evolved novel ways to protect their 5’ ends

Because these elements transpose onto chromosome ends, their 5’ DNA is exposed to terminal erosion until another element transposes onto the chromosome and becomes the new end. As a result, telomere arrays contain elements with varying amounts of 5’ truncation. Unlike telomerase-maintained telomeres, these arrays must also provide a breeding stock of elements capable of transposing to extend chromosome ends. Therefore, terminal telomeric elements must occasionally escape loss of essential 5’ sequence, and we found that they have evolved novel ways to do just that.

The first evidence for novel 5’ end protection came from our study of HeT-A promoters in D. melanogaster. We found that his non-LTR element had an unusual promoter resembling an evolutionary precursor of an LTR element (or retrovirus). The HeT-A promoter lies at the 3´ end of the element and promotes transcription of the adjacent downstream element (if it too is HeT-A) rather than the element in which it resides. Thus the end of the upstream element is acting as a 5’LTR (Long Terminal Repeat) for the element that is transcribed. Because it actually begins in this “5’LTR”, transcription adds a small amount of non-essential sequence to the 5’end of the RNA (which we call a Tag). This Tag forms the extreme end of the telomere when the RNA is reverse transcribed onto a chromosome and is subject to small scale erosion until another element transposes onto it. In telomere arrays we can read the history of this non-essential sequence because varying amounts of Tag sequence are found at the 5’ends of complete elements, evidence that a new transposition had capped the end before the Tag had completely eroded. Furthermore, when an element with a partial Tag is transcribed a new complete Tag is tacked on the RNA’s 5’end, which will then be subject to erosion when it transposes. Thus elements can amass strings of variously eroded Tags, evidence that an element has transposed multiple times. The history recorded in Tag lengths shows that these sequences are adequate to protect the 5’ends of many elements).

The unusual HeT-A promoter provides an evolutionary advantage for an element serving as a telomere. To see whether this advantage was conserved, we analyzed HeT-A’s partner, TART, in D. melanogaster and their homologues in the distantly related D. virilis. (TAHRE is too rare to study). D. virilis TART (TARTvir ) has the HeT-Amel promoter and Tags. Surprisingly, neither HeT-Amel’s partner, TARTmel , nor its homologue, HeT-Avir, has this promoter. TARTmel and HeT-Avir each has a typical non-LTR promoter. These typical promoters are in the 5’UTR (untranslated region) of the element, immediately downstream of the transcription start site so that the promoter is included in the RNA transposed to the new site. We discovered that, in spite of having a conventional promoter, TARTmel has a very unconventional method of adding non-essential sequence to its 5’ end. Our experimental promoter analyses and mapping of transcription start sites, combined with extensive database searches and sequence analyses allow us to propose a model, similar in principle but not in detail, to the concerted evolution of retroviral LTRs. In this model, when TARTmel RNA is reverse transcribed onto the end of the chromosome, the reverse transcriptase makes a template jump from the 5’ end of the RNA back to sequence in the 3’ end of the RNA and recopies part of this to extend the 5’ end of the element with redundant sequence. HeT-Avir is still an enigma. It must have a mechanism to protect its 5’ end because we have seen intact elements in telomere clones but the 5’ ends have no obvious added sequence.

It is intriguing that these mechanisms of 5’end protection seem to be somewhat variable on an evolutionary time scale because only two of the four elements share the same mechanism. However, only transposition-competent elements can give rise to lineages of new elements, providing a strong drive for evolving efficient 5’end protection. The diversity of mechanisms we have found seems to be a result of this drive.

Differential maintenance of sequence in telomeric and centromeric heterochromatin

Transposable elements have significantly added to and shaped eukaryotic genomes. Unfortunately, difficulties in assembling long sequences of highly repetitive DNA have largely precluded analysis of the organization and possible roles of transposable elements in heterochromatic regions such as centromeres and telomeres. We were fortunate that sequence of a BAC (Bacterial Artificial Chromosome) containing sequence from the centromere of the D. melanogaster Y chromosome was published recently. This BAC contains an array of telomeric HeT-A elements which apparently transposed from a telomere into the Y centromere more than 13 Myr ago. The transposed array remained there as the genus evolved and is now found in all species in the D. melanogaster group of species. Our comparison of this sequence with telomere array sequences provides the first evidence that a segment of DNA can be maintained very differently in centromeric heterochromatin and telomeric heterochromatin .

The differential maintenance of the HeT-A arrays results in sequence organizations that reflect different roles in the two chromosomal environments. The telomere array has grown only by transposition of new elements to the chromosome end; whereas the centromeric array has grown by repeated amplification of segments of the original telomere array. In the telomere, many elements have undergone variable 5’-truncation by gradual erosion and irregular deletion of the chromosome end; however a significant fraction remain complete and capable of further retrotransposition. In contrast, each element in the centromere region has lost 40% or more of its sequence and the loss occurred by internal, rather than terminal, deletions. No centromere element retains a significant part of its original coding region. Thus the centromeric array has been restructured to resemble the highly repetitive satellite sequences typical of centromeres in multicellular organisms; meanwhile, over a similar or longer time period, the telomere array has maintained its ability to provide retrotransposons competent to extend telomere ends.


Pardue M-L and DeBaryshe PG (2011) Retrotransposons that maintain chromosome ends. Proc. Natl. Acad. Sci. U S A. 108(51):20317-24. www.pnas.org/cgi/doi/10.1073/pnas.1100278108

Pardue M-L and DeBaryshe PG (2011) Adapting to life at the end of the line. How Drosophila telomeric retrotransposons cope with their job. Mobile Genetic Elements 1(2):128-134. Epub ahead of print. http://dx.doi.org/10.4161/mge.1.2.16914

DeBaryshe PG and Pardue M-L (2011) Differential maintenance of retrotransposon DNA in Drosophila telomeric and centromeric heterochromatin. Genetics 187(1):51-60. Epub 2010 Nov 1. PMC3018307 21041555[PMID]

George JA, Traverse KL, DeBaryshe PG, Kelley KJ and Pardue M-L (2010) Evolution of diverse mechanisms for protecting chromosome ends by the Drosophila TART telomere retrotransposons. Proc. Natl. Acad. Sci. USA 107(49):21052-7. Epub 2010 Nov 18. PMC3000255 21088221[PMID]

Traverse KL, George JA, DeBaryshe PG and Pardue M-L (2010) Evolution of species-specific promoter-associated mechanisms for protecting chromosome ends by Drosophila Het-A telomeric transposons. Proc. Natl. Acad. Sci. USA 107(11):5064-5069. Epub 2010 Mar 1. PMC2841908 20194755[PMID]

Fuller AM, Cook EG, Kelley KJ and Pardue M-L (2010) Gag proteins of Drosophila telomeric retrotransposons: collaborative targeting to chromosome ends. Genetics 184(3):629-636. Epub 2009 Dec 21. PMC2845333 20026680[PMID]

Free PMC Article

Pardue M-L and DeBaryshe PG (2008) Drosophila telomeres: a variation on the telomerase theme. Fly 2(3):101-110. 18820466[PMID]

Casacuberta E, Azorín MarínF, and Pardue M-L (2007) Intracellular targeting of telomeric retrotransposon Gag proteins of distantly related Drosophila species. Proc. Natl. Acad. Sci, USA 104:8391-8396. PMC1895960 17483480[PMID]

George JA, DeBaryshe PG, Traverse KL, Celniker SE, and Pardue M-L (2006) Genomic organization of the Drosophila telomere retrotransposable elements. Genome Research 16:1231-1240. Epub 2006 Sep 8. PMC1581432 16963706[PMID]

Rashkova S, Athanasiadis A, and Pardue M-L (2003) Intracellular targeting of Gag proteins of the Drosophila telomeric retrotransposons. J. Virol. 77:6376-6384. PMC155015 2743295[PMID]

Pardue M-L and DeBaryshe, PG (2003) Retrotransposons provide an evolutionarily robust non-telomerase mechanism to maintain telomeres. Annu. Rev. Genet.37:485-511. Epub 2003 Aug 6. 14616071[PMID]