The Zakian Lab uses a combination of genetic and biochemical methods to study replication fork progression and telomere structure/replication mainly in budding and fission yeasts with occasional forays into mammalian systems. Major accomplishments include:
Ciliate termini and YACs:The Zakian Lab were the first to construct and characterize a linear artificial chromosome (Dani and Zakian 1983 PNAS). In this and subsequent work (Pluta et al. 1984 PNAS), they were one of two groups to use ciliate telomeres to generate linear yeast episomes, a strategy that began the molecular era of yeast telomere biology.
Single strand telomere binding proteins:The Lab isolated and characterized the first single-strand, sequence specific telomere binding protein, the prototype of Pot1, from ciliates (Gottschling and Zakian 1986 Cell). In vitro, the heterodimeric complex is sufficient to distinguish authentic telomeres from DNA ends and to protect them from degradation. They were one of two groups to demonstrate that Cdc13 is the yeast functional equivalent of the ciliate G-strand binding protein (Lin and Zakian 1996 PNAS) and the first to show that it is telomere associated in vivo (Bourns et al. 1998 Mol Cell Biol; Tsukamoto 2001 Current Biol). Using genetic and biochemical approaches, they demonstrated that Cdc13 and the telomerase subunit Est1 interact directly, an interaction that is sufficient to bring telomerase to DNA ends in vitro (Qi and Zakian 2000 Genes Dev; Wu and Zakian 2011 PNAS).
Duplex telomere binding proteinsThe Zakian lab was the first to show that the duplex sequence specific Rap1 protein binds telomeres in vivo where it affects telomere structure and chromosome stability (Conrad et al. 1990 Cell). These studies were the first chromatin immunoprecipitation (ChIP) of a yeast protein. They extended this work by showing that yeast telomeric DNA is assembled into a non-nucleosomal chromatin structure (Wright et al. 1992 Genes Dev).
Telomere position effects:The Zakian lab discovered telomere position effect, TPE, the transcriptional repression of genes near telomeres in budding yeast (Gottschling et al., 1990Cell). We showed that silencing is due to telomeric sequence, not to its position at chromosome ends (Stavenhagen and Zakian 1994Genes & Develop.) and that telomeric DNA also exerts position effects on recombination (Stavenhagen and Zakian 1998Genes & Develop.). Yeast TPE requires the ability of the telomere to fold back onto subtelomeric DNA (“telomere folding”; de Bruin et al., 2000Mol. Cell. Biol). The Zakian lab was the first to use the LacO-GFP system to localize individual telomeres within the nucleus (Tham et al. 2001Molec Cell). Although the VII-L telomere is often at the nuclear periphery, this localization does not require key silencing proteins. Thus, localization and TPE are not obligatorily linked.
Telomere structure:By sequencing individual telomeres, the Zakian lab showed that in wild type cells, only the terminal half of yeast telomeres is subject to lengthening and degradation (Wang and Zakian 1990 Mol Cell Biol). They made the surprising discovery that C-strand degradation is a cell cycle regulated step in telomere biology that allows G-tails to be generated even in telomerase deficient cells (Wellinger et al. 1993 Cell; Wellinger et al. 1996 Cell). This degradation occurs after conventional semi-conservative telomere replication, while C-strand resynthesis occurs prior to mitosis (Wellinger et al. 1993 Mol Cell Biol).
In addition to interacting with Est1, Cdc13 also interacts with the catalytic subunit of DNA polymerase alpha (Qi and Zakian 2000 Genes Dev). When this interaction is reduced, telomerase is more likely to lengthen telomeres. These data were the first evidence for Cdc13 acting as a switch between promoting telomerase lengthening versus C-strand resynthesis. By eliminating a single telomere in a controlled manner, the Zakian lab demonstrated that while telomeres are essential for the long term stability of yeast chromosomes, a chromosome without a telomere can replicate and segregate for many generations before it is lost (Sandell and Zakian 1993 Cell). During this analysis, it was also revealed that even a single double strand break generates a robust checkpoint response and discovered adaptation, the ability of cells to recover from arrest even in the presence of an unrepaired double strand break.
Telomere recombination:There are two major pathways for telomere maintenance, telomerase and recombination. the Zakian lab provided the first evidence that eukaryotic telomeres can be lengthened by gene conversion (Pluta and Zakian 1989 Nature; Wang and Zakian 1990 Nature). the Zakian lab also showed that type II recombination results in very heterogenous length telomeres that are generated by abrupt but relatively infrequent Rad50 dependent, Rif2 inhibited telomere lengthening followed by gradual loss of telomeric DNA (Teng and Zakian 1999 Mol Cell Biol; Teng et al. 2000 Molec Cell). These events are remarkably similar to the telomerase-independent ALT pathway of mammals.
Cell cycle regulation of budding yeast telomerase:The Zakian lab pioneered the use of ChIP to study the dynamics of telomeric chromatin as a function of the cell cycle. Although yeast telomerase is not active in G1 phase, the catalytic core of telomerase is telomere associated at this time. However, telomerase subunits Est1 (Taggart et al., 2002 Science) and Est3 (Tuzon et al. 2011PLoS Genetics) are telomere bound only in late S phase when telomerase is active. This late S phase association requires specific interactions between Est1 and both Cdc13 and TLC1 RNA (Chan et al. 2008PLoS Genetics) as well as between Est1 and Est3. As purified Est1 interacts with both Cdc13 and Est3in vitro, the protein-protein interactions needed to bring telomerase to telomeres are likely direct (Tuzon, et al., 2011PLoS Genetics; Wu and Zakian 2011PNAS). Est1 is the only core telomerase subunit whose abundance is cell cycle regulated (Taggart et al., 2002Science). Regulation of Est1 levels and hence of telomere length depend on the trimeric Cdc48-Npl4-Ufd1 complex, which targets ubiquinated Est1 for proteosome-mediated degradation (Lin et al., under review). The Cdc48 complex, which is telomerase-associated in both G1 and G2/M phase, can thus prevent premature assembly of telomerase in G1 phase and disassembly of telomerase at the end of the cell cycle.
Preferential targeting of telomerase to short telomeres:We applied our telomere ChIP assay to understand how telomerase targets short telomeres for lengthening. Est1 and Est2 bind preferentially to short telomeres in late S phase, which can explain their preferential lengthening by telomerase (Sabourin et al, 2007Molec. Cell). Positive and negative factors are responsible for establishing this preference. Short telomeres bind higher levels of the trimeric Mre11 complex, which in turn, recruits the Tel1 kinase preferentially to short telomeres (Goudsouzian et al 2006Molec. Cell;Sabourin et al., 2007Molec. Cell; McGee et al., 2010Nature Struct. Mol. Bio). High Tel1 binding acts redundantly with Tbf1 to channel telomerase to short telomeres. Three proteins bind more robustly to long telomeres: Rap1, Rif2 (which interacts with Rap1) and the Pif1 DNA helicase (Sabourin et al., 2007Molec. Cell;). Reduced Rif2 at short telomeres is required for high telomerase binding (McGee et al. 2010Nature Struct. Mol. Bio). In the absence of Pif1, telomerase no longer binds preferentially to short telomeres (Phillips et al., 2015PLoS Genetics).
Regulation of fission yeast telomerase: We also used ChIP to determine protein dynamics atS. pombetelomeres. As a prelude to the ChIP experiments, we isolated and characterized the illusiveS. pombetelomerase RNA (Webb and Zakian 2008Nature Struct. Mol. Biol). We showed that the conserved stem terminus element (STE) in telomerase RNA is essential for telomerase activityin vivoandin vitro(as it is in humans) (Webb and Zakian, under review). Surprisingly, the STE, which is far from the template region of the RNA in the primary sequence, affects the sequence of telomeric DNA by controlling the efficiency of the template boundary element. As a result of its effect on telomere sequence, a wild type STE is required to maintain correct shelterin binding. Recruitment of fission yeast telomerase to telomeres requires interaction of the telomere bound shelterin protein, Ccq1, with Est1 (Webb and Zakian 2012Genes & Develop.).
The Pif1 DNA helicase inhibits telomerase:TheS. cerevisiaePif1 DNA helicase uses its helicase activity to suppress telomerase-mediated telomere elongation andde novotelomere addition to double strand breaks (Schulz and Zakian 1994Cell; Zhou et al., 2000Science). Telomere addition is a particularly dangerous event as it results in loss of heterozygosity for all sequences distal to the break. By bothin vivoandin vitroassays, Pif1 uses its ATPase activity to evict telomerase from DNA (Boulé et al., 2005Nature). Pif1 has the unusual property of preferentially displacing RNA from DNA, suggesting that it dislodges telomerase by unwinding the telomerase RNA/telomeric DNA hybrid (Boulé and Zakian 2005NAR; Zhou et al., 2014eLIFE). Pif1 reduces both telomerase processivity and the fraction of telomeres that are lengthened by telomerase in a given S phase (Phillips et al., 2015PLoS Genetics). Another yeast helicase, Hrq1 acts non-catalytically to inhibit telomerase at both telomeres and DNA breaks. Hrq1 is also required for repair of DNA inter-strand cross-links (Bochman et al., 2014Cell Reports). Hrq1 is a homolog of human RecQ4 whose mutation causes increased susceptibility to cancer and premature aging,
Rrm3 and Pfh1 DNA helicases promote fork progression through stable protein complexes:Virtually all eukaryotes and many bacteria encode Pif1 family DNA helicases (Zhou et al, 2000Science; Paeschke et al., 2013Nature). Some, like budding yeast, encode two (Pif1 and Rrm3) while most organisms, like fission yeast (Pfh1) and humans (hPIF1) encode only one. Rrm3 promotes fork progression at over 1000 genomic loci, including ribosomal DNA (Ivessa et al., 2000Cell), telomeres (Ivessa et al., 2002Genes & Develop.), polymerase III transcribed genes, centromeres, silencers, and converged replication forks (Ivessa et al., 2003Molec. Cell; Azvolinsky et al., 2009Molec. Cell; Fachinetti et al., 2010Molec. Cell). Rrm3-sensitive sites are assembled into stable protein complexes, and it is these complexes that make their replication Rrm3-dependent (Ivessa et al., 2003Molec. Cell; Torres et al., 2004Genes & Develop.). TheS. pombePfh1 helicase also promotes fork progression past stable protein complexes (Sabouri et al., 2012Genes & Develop.). Pif1 family helicases are the first examples of eukaryotic helicases that promote fork progression at multiple classes of hard-to-replicate sites, such as protein complexes, highly transcribed genes, and DNA secondary structures (see below). As a result of genome-wide studies to identify new Rrm3-sensitive sites, we discovered that highly transcribed genes impede fork progression in an orientation independent method, the first demonstration of what is now known to be a general phenomenon (Azvolinky et al., 2009Molec. Cell). In another study relevant to fork progression, we found that tri-nucleotide repeats cause length dependent replication fork slowing and DNA breakage in budding yeast and implicated Rad27 (homolog of human FEN1) in suppressing repeat expansion (Freudenreich et al. 1998Science).
Pif1 family helicases from bacteria to humans unwind G-quadruplex (G4) DNAin vitroand/or suppress G4-induced DNA instabilityin vivo:G4 DNA is an extremely stable DNA secondary structure held together by G-G base pairs. Sequences that can form G4 DNA, called G4 motifs, are found throughout bacterial and eukaryotic chromosomes. In budding and fission yeasts, the position and sequence of G4 motifs are evolutionarily conserved and associated with distinct genomic features, such as promoters, UTRs, ribosomal DNA, telomeres, etc (Capra et al., 2010PLoS Computer Sci; Sabouri et al. 2014BMC Biology). In both budding and fission yeasts, G4 motifs are among the preferred binding sites for their respective Pif1 family helicases (i.e., Pif1 in budding and Pfh1 in fission yeasts) (Pasechke et al., 2011Cell; Sabouri et al, 2015BMC Biology). In the absence of Pif1/Pfh1, DNA replication slows and breakage often occurs at the G4 motifs that are helicase-associated in wild type cells. G4-associated DNA damage was quantified in budding yeast using a gross chromosomal rearrangement (GCR) assay (Paeschke et al., 2013Nature). Although G4 motifs do not cause GCR events in wild type cells, novel types of GCR events were associated with replication through G4 motifs in the absence of Pif1 and/or Rrm3. Bacterial,S. pombe, and human Pif1 enzymes can suppress G4-associated DNA damage in Pif1-deficientS. cerevisiae. Budding yeast and bacterial Pif1 helicases unwind G4 structuresin vitrowith remarkable ease. Our ensemble biochemistry on Pif1 helicases was confirmed and extended by single molecule analyses done in collaboration with T.J. Ha (Zhou et al. 2014eLIFE). These single molecule assays also confirmed that Pif1 is particularly active at unwinding RNA/DNA hybrids.
Mammalian Pif1 helicases:Our early work on mouse and human PIF1 helicases was not particularly informative, although it did suggest a role forthese helicases at telomeres (Mateyak and Zakian 2006Cell Cycle.; Snow et al., 2007Mol. Cell. Biol.). Using budding yeast as a “test tube”, hPIF1 inhibits telomerase and suppresses G4-induced instability when expressed in Pif1-deficientS. cerevisiae, suggesting that it might also do so in its natural context. In collaboration with Mary Claire King and co-workers, we reported that multiple families with increased breast cancer risk have a mutation in the most conserved portion of hPIF1, the Pif1 signature motif, and the analogous mutation in fission yeast Pfh1 does not provide Pfh1 functions inS. pombe(Chisholm et al., 2012PLoS One). We are in the process of making conditional hPIF1 knockout/mutant human cell lines to determine if hPIF1 has roles similar to those of its yeast relatives. hPIF1 likely has important functions as the doubling time of a heterozygous cell line (hPIF1-KO/WT) is twice as long as that of WT cells (C. Follonier and VAZ, unpublished).