Sir

T-loop formation and abrupt telomere shortening

There are several important issues in A. J. Lustig's recent review1 that I feel should be addressed. First, the title of the review ('Clues to catastrophic telomere loss in mammals from yeast telomere rapid deletion') indicates that conclusions should be drawn about catastrophic telomere loss in mammals from studies in yeast. However, the entire discussion is based on results that were obtained from telomere studies in mammals, and in particular, t-loop formation and deletion events, which have not yet been identified in yeast and are still speculative1. Second, the whole review is based on a molecular model for abrupt telomere shortening (ATS) in mammals. However, Lustig fails to mention that this model of t-loop formation and telomere deletion through self-recombination in human cells had already been described in a previous publication2, which indicated that it might be central to the ageing process.

Following Lustig's publication and other publications in the field, it is evident that there is a serious lack of information about the origin of the molecular model for t-loop formation and catastrophic telomere loss in normal and immortal mammalian cells. Abrupt telomere shortening as a biological phenomenon has been described both in yeast during studies of the telomere-binding protein Rap13 and in human SV40-immortalized cells4. In later studies, by following the dynamics of individual, marked telomeres on chromosome 13, sudden and stochastic changes in telomere lengths were clearly observed4. On the basis of these and other findings, we put forward a molecular model2 that could explain not only these phenomena, but also some important features of normal cell growth, such as the stochastic nature of cellular ageing5. The model is based on both gradual and abrupt telomere shortening. The main contributor to telomere erosion is progressive terminal shortening owing to the end-replication problem6, superimposed with occasional catastrophic telomere deletion as a result of intrachromatid recombination. This event is initiated by strand invasion of the 3′ overhang into the telomeric/subtelomeric border region followed by formation of a t-loop, which was later identified in normal and immortal human and mouse cells7. Our model also predicts that long telomeres have a low probability of undergoing such abrupt shortening, but, as shortening progresses, this probability increases almost exponentially2. When Lustig earlier used the t-loop model to explain telomere rapid deletion (TRD) in yeast in 2001 (Ref. 8), all of the important features of the ATS model had already been examined by electron microscopy7 and cytogenetically9. Today, we can say that all of the predictions of our original model have been examined both structurally and functionally: t-loop formation7, recombination proteins at the t-loop9, changes in telomere structure with shortening10, ATS deletion product, telomeric circular DNA11 and finally, the telomere-deletion event in normal cells during senescence12. The idea that catastrophic telomere loss happens in mammalian cells was also proposed in our earlier papers2,13.

In summary, our original theoretical presentation of a t-loop-mediated telomere-deletion model has since been borne out by numerous experimental data, yet many interesting aspects of these processes remain to be studied.