Our findings suggest a dynamic cross talk between TRF1 and TRF2 and provide a molecular mechanism for telomere length homeostasis by TRF2 in the absence of TRF1.

Determining mechanisms of drug action in human cells remains a major challenge. Here we describe an approach in which multiple drug resistant clones are isolated and transcriptome sequencing is used to find mutations in each clone. Further analysis of mutations common to more than one clone can identify a drug’s physiological target and indirect resistance mechanisms, as indicated by our proof-of-concept studies of cytotoxic anti-cancer drugs, BI 2536 and bortezomib.

Previous studies in human cells indicate that sister telomeres have distinct requirements for their separation at mitosis. In cells depleted for tankyrase 1, a telomeric poly(ADP-ribose) polymerase, sister chromatid arms and centromeres separate normally, but telomeres remain associated and cells arrest in mitosis. Here, we use biochemical and genetic approaches to identify proteins that might mediate the persistent association at sister telomeres. We use immunoprecipitation analysis to show that the telomeric proteins, TRF1 (an acceptor of PARsylation by tankyrase 1) and TIN2 (a TRF1 binding partner) each bind to the SA1 ortholog of the cohesin Scc3 subunit. Sucrose gradient sedimentation shows that TRF1 cosediments with the SA1-cohesin complex. Depletion of the SA1 cohesin subunit or the telomeric proteins (TRF1 and TIN2) restores the normal resolution of sister telomeres in mitosis in tankyrase 1-depleted cells. Moreover, depletion of TRF1 and TIN2 or SA1 abrogates the requirement for tankyrase 1 in mitotic progression. Our studies indicate that sister telomere association in human cells is mediated by a novel association between a cohesin subunit and components of telomeric chromatin.

Human telomere function is mediated by shelterin, a six-subunit complex that is required for telomere replication, protection, and cohesion. TIN2, the central component of shelterin, has binding sites to three subunits: TRF1, TRF2, and TPP1. Here we identify a fourth partner, heterochromatin protein 1g (HP1g), that binds to a conserved canonical HP1-binding motif, PXVXL, in the C-terminal domain of TIN2. We show that HP1g localizes to telomeres in S phase, where it is required to establish/maintain cohesion. We further demonstrate that the HP1-binding site in TIN2 is required for sister telomere cohesion and can impact telomere length maintenance by telomerase. Remarkably, the PTVML HP1-binding site is embedded in the recently identified cluster of mutations in TIN2 that gives rise to dyskeratosis congenita (DC), an inherited bone marrow failure syndrome caused by defects in telomere maintenance. We show that DC-associated mutations in TIN2 abrogate binding to HP1g and that DC patient cells are defective in sister telomere cohesion. Our data indicate a novel requirement for HP1g in the establishment/maintenance of cohesion at human telomeres and, furthermore, may provide insight into the mechanism of pathogenesis in TIN2-mediated DC.

Structure of the meiotic spindle T he meiotic spindle is made up of shorter microtubules than previously believed, suggest results (Rockefeller University, New York, NY), and colleagues. Current models of the spindle, as a bipolar array of overlapping fi laments extending from opposite spindle poles, will require revision. To get a closer look at the architecture of the meiotic spindle, Yang et al. incorporated labeled tubulin subunits into the spindle in a cell-free system. By refi ning their fl uorescent speckle microscopy techniques, the authors were able for the fi rst time to track individual tubulin subunits (seen as speckles) in a single tubulin polymer. The authors identifi ed pairs of speckles representing subunits on the same fi lament. Speckle separation supplied them with the minimum length of that fi lament. They then fi tted a mathematical model to these observed lengths to predict overall fi lament lengths: most fi laments were only 40% of the total spindle length. The short fi laments were also scattered throughout the spindle. The researchers now propose that the spindle is a tiled array of overlapping short fi laments. The group next examined how spindle-associated proteins might control fi lament and spindle size. By inhibiting microtubule motor proteins, they found that dynein-dynactin limited individual fi ber lengths and thus overall spindle length. Kinesin 5 activity limited the overlap between fi bers by sliding them apart. "Our work suggests the spindle is a self-organizing system, whose stability and functional characteristics are built on these kind of local interactions," says Kapoor. Localized mRNA is the norm L ocation, location, location. It's critical for real estate, proteins, and-according to work by Eric Lécuyer, Henry Krause, and colleagues (University of Toronto, Canada)-mRNAs, too. Several localized mRNAs have been previously studied, but just how many transcripts are localized in the cell, and in what patterns, is unknown. Lécuyer et al. approached this problem by optimizing fl uorescence in situ hybridization (FISH) in a global analysis of developmentally expressed mRNAs. They found that 71% of the mRNAs in early fl y embryos showed specifi c patterns of subcellular localization. In several cases, they found new examples of mRNAs that colocalized with their protein products. Less energy is probably required to transport a few copies of an mRNA than to move around many more copies of the protein. And the proteins will be created where they are needed and possibly prevented from straying where they are not wanted. "We need to revise the textbook image of proteins being made in a centralized location near the nucleus, then traffi cking to their ultimate locations," says Krause. "Our work shows that the mRNAs are an intelligent actor, not just a dumb vehicle for creating proteins." With their new database, the group can now further investigate how and why mRNAs are localized.