TZAP‐ing telomeres down to size

Abstract

The phenomenon of gradual telomere shortening has become a paradigm for how we understand the biology of aging and cancer. Cell proliferation is accompanied by cumulative telomere loss, and the aged cell either senesces, dies or transforms toward cancer. This transformation requires the activation of telomere elongation mechanisms in order to restore telomere length such that cell death or senescence programs are not induced. Most of the time, this occurs through telomerase reactivation. In other rare cases, the Alternative lengthening of telomeres (ALT) pathway hijacks DNA recombination‐associated mechanisms to hyperextend telomeres, often to more than 50 kb. Why telomere length is restricted and what sets their maximal length has been a long‐standing puzzle in cell biology. Two recent studies published in this issue of EMBO Reports [1] and recently in Science [2] sought to address this important question. Both built on omics approaches that identified ZBTB48 as a potential telomere‐associated protein and reveal it to be a critical regulator of telomere length homeostasis by the telomere trimming mechanism. These discoveries provide fundamental insights for our understanding of telomere trimming and how it impacts telomere integrity in stem and cancer cells.

ZBTB48 is a Krüppel‐like C2H2 zinc finger protein consisting of 11 tandem zinc finger domains located C‐terminal to the BTB/POZ domain. Both studies establish that ZBTB48, renamed by Li et al 2 as TZAP for telomeric zinc finger‐associated protein, has a greater tendency to bind hyperextended telomeres in mouse stem cells and cancer cells, irrespective of whether telomerase or ALT is operational. Jahn et al 1 note that TZAP could in some instances bind to shorter telomeres. By manipulating TZAP's coding sequence, both demonstrate that TZAP binds to telomeric DNA largely via the three C‐terminal zinc fingers 9–11 without contacting or bridging through TRF1 and TRF2—core components of the canonical telomeric DNA binding complex shelterin. Using their expertise in genetic manipulation of telomere structure, Li et al 2 elegantly demonstrate that fusing these terminal zinc finger domains of TZAP with a fragment of TRF2 that is dispensable for binding to telomeric DNA prevents the fusion of chromosomes via uncapped telomeres that occurs when TRF2 is deleted. This finding indicates that the mere presence of protein bound to telomeric DNA can foster chromosome end protection.

The two studies (…) break new ground by revealing that the process of telomere trimming relies on TZAP

Based on the preferential association of TZAP with longer telomeres, Li et al 2 postulate that as telomeres extend beyond normal limits of homeostasis, the occupancy of TRF2 decreases, thereby enabling TZAP binding. Indeed, they show that flooding cells with TRF2 reduces the association of TZAP to telomeres. Therefore, TZAP appears to compete with TRF1 and TRF2 for binding to telomeres, and its association could be contingent on the exhaustion of free shelterin, possibly as telomeres replicate and elongate (Fig 1, top panel). This is consistent with evidence that excessive telomere elongation is counter‐acted by shelterin (TRF1, TRF2) 3. It also suggests that the supply of TRF1 and TRF2 protein is tightly regulated and incapable of preventing excessive telomere elongation once the upper limits are reached whereupon a second, more enigmatic, mechanism of telomere homeostasis termed telomere trimming is activated 4 (Fig 1, middle panel).

Figure 1. Control of telomere length by TZAP.

In normal cells, telomere length is regulated, in part, through the high density of shelterin (TRF1/TRF2) that blocks TZAP binding. Once the normal limits of telomere length are breached and shelterin occupancy is reduced as happens in cancer cells (particular ALT cells), TZAP binds to telomeric DNA in order to trim it and prevent excessive extension of telomeres. This process, telomere trimming, generates extra‐chromosomal telomeric DNA species like double‐stranded T‐circles (TC) and partially single stranded C‐circles (CC). Genetic manipulation to delete TZAP in ES cells promotes further extension of telomeres and may be linked with genomic instability that is potentially wrought by further diminution of shelterin binding across hyperextended telomeres. TZAP deletion simultaneously reduces expression of MTFP1 and thereby negatively impacts mitochondrial integrity. How this apparent cross talk is regulated and whether the same occurs in stem cells remain to be determined but TZAP could affect genomic stability in multiple ways.

The two studies by Li et al 2 and Jahn et al 1 break new ground by revealing that the process of telomere trimming relies on TZAP. First, they show that disrupting TZAP by CRISPR/Cas9 gene targeting in mouse embryonic stem cells and cancer cells leads to significant elongation of telomeres. Although they differ on whether TZAP deletion leads to lengthening of ALT telomeres, both conclusively show that TZAP deletion reduces the abundance of extra‐chromosomal telomeric (ECT)‐DNA species termed C‐circles and T‐circles that are surrogate markers of ALT activity but have also been linked with telomere trimming in normal cells 4, 5. Furthermore, Li et al 2 reveal that overexpressing TZAP stimulates ECT‐DNA production—which is consistent with a potential trimming activity of TZAP. The generation of C‐circles, which contain single‐stranded C‐rich telomeric DNA, has been linked to active synthesis of telomeric DNA, especially in response to induction of the ALT phenotype by replicative stress or FokI‐induced double‐strand breaks 6. The latter appears to rely on the assembly of a dedicated replisome that extends ALT telomeres outside the restrictions of S‐phase 6. This raises intriguing questions as to TZAP's relationship with the polymerases that synthesize telomeric DNA in both normal and ALT cancer cells, and whether TZAP could associate directly or proximally to the replisome. On the other hand, T‐circles arise through XRCC3‐dependent resolution of telomere recombination intermediates that generate large deletions of telomeric DNA 5. Although TZAP itself does not harbor any obvious nucleolytic or enzymatic activity, the multifaceted SLX4‐MUS81‐XPF nucleolytic complex has been shown to associate with hyperextended telomeres and could work in conjunction with TZAP to nick, cleave, or resolve such DNA intermediates as part of the trimming mechanism. Thus, there is much to be determined as to how and in which context TZAP‐associated trimming proceeds. Deciphering whether TZAP has a particular affinity for structured DNAs and its network of protein interactions should provide essential insights. The role of TZAP's POZ domain could be key since POZ domains can serve as platforms of protein–protein interactions and protein modification 7.

Precisely why trimming is activated is unclear but it is thought to prevent potential genomic destabilization by hyperextension of telomeres—a notion recently substantiated by the recognition of hyperextended telomeres in human embryonic stem cells as sources of DNA damage 8 (Fig 1, bottom panel). That trimming has been detected in germ 5 and stem cells 6 suggests that it could be consequential for early development and differentiation. The manipulation of TZAP offers the possibility to address this question through the development of mouse and stem cell TZAP‐deficient models.

Complicating the issue is the discovery by Jahn et al 1 that TZAP also moonlights as a transcription factor. Rather than representing a more general transcriptional regulator, TZAP localizes to a small, defined number of promoters that are somewhat repressed in its absence. What is truly remarkable is that not only does TZAP localize to the promoter of the mitochondrial fission protein, MTFP1, but cells lacking TZAP completely lose expression of MTFP1 protein and display defects in mitochondrial biogenesis (Fig 1, bottom panel). Mitochondrial‐telomere cross talk has been described in relation to dysfunction of either, leading to alterations in mitochondrial metabolite production that impact nuclear epigenetic modification and DNA repair activities that ultimately feedback and impinge on their structural integrity 9, 10. Thus, TZAP could conceivably influence either. As mentioned above, deciphering its protein interactions will be key and could perhaps allow for the development of separation of function alleles to unravel this compelling question.

See also: A Jahn et al (June 2017) and JSZ Li et al (February 2017)