Dynamics of double-strand break repair


In a new paper in the journal Genetics, former Brandeis postdoc Eric Coïc and undergrads Taehyun Ryu and Sue Yen Tay from Professor of Biology Jim Haber’s lab, along with grad student Joshua Martin and Professor of Physics Jané Kondev, tackle the problem of understanding the dynamics of homologous recombination after double strand breaks in yeast. According to Haber,

The accurate repair of chromosome breaks is an essential process that prevents cells from undergoing gross chromosomal rearrangements that are the hallmark of most cancer cells.  We know a lot about how such breaks are repaired.  The ends of the break are resected and provide a platform for the assembly of many copies of the key recombination protein, Rad51.  Somehow the Rad51 filament is then able to facilitate a search of the entire DNA of the nucleus to locate identical or nearly identical (homologous) sequences so that the broken end can pair up with this template and initiate local copying of this segment to patch up the chromosome break.  How this search takes place remains poorly understood.

The switching of budding yeast mating type genes has been a valuable model system in which to study the molecular events of broken chromosome repair, in real time.  It is possible to induce synchronously a site-specific double-strand break (DSB) on one chromosome, within the mating-type (MAT) locus.  At opposite ends of the same chromosome are two competing donor sequences with which the broken ends of the MAT sequence can pair up and copy new mating-type sequences into the MAT locus.

Normally one of these donors is used 9 times more often than the other.  We asked if this preference was irrevocable or if the bias could be changed by making the “wrong” donor more attractive – in this case by adding more sequences to that donor so that it shared more and more homology with the broken ends at MAT.  We found that the competition could indeed be changed and that adding more homologous sequences to the poorly-used donor increased its use.


In collaboration with Jané Kondev’s lab we devised both a “toy” model and a more rigorous thermodynamic model to explain these results.  They suggest that the Rad51 filament carrying the broken end of the MAT locus collides on average 4 times before with the preferred donor region before it actually succeeds in carrying out the next steps in the process that lead to repair and MAT switching.

Dynamics of homology searching during gene conversion in Saccharomyces cerevisiae revealed by donor competition Eric Coïc , Joshua Martin, Taehyun Ryu, Sue Yen Tay, Jané Kondev and James E. Haber. Genetics. 2011 Sep 27 2011 Sep 27

Formins require assistance; not so different from other actin nucleators

Formins are a family of proteins conserved across a wide range of eukaryotes and constitute a major class of actin nucleators. In a paper recently published in Molecular Biology of the Cell, a team led by Ph.D. student Brian Graziano in the laboratory of Professor Bruce Goode made the surprising finding that formins depend on co-factors to efficiently nucleate actin assembly both in vitro and in vivo. This discovery was unanticipated because earlier studies had shown that purified formins are sufficient to catalyze actin polymerization in vitro. Graziano, working in collaboration with the labs of Laurent Blanchoin and Isabelle Sagot, investigated the mechanism and function of a formin-binding protein called Bud6 and found that it elevates formin nucleation activity by 5-10 fold. Further, they showed that this activity of Bud6 is critical in vivo for maintaining normal levels of actin cable assembly and polarized cell growth (see figure).

Earlier work from the Goode lab had shown that Bud6 enhances formin-mediated actin assembly in vitro (Moseley et al., 2004), but had left open the question of whether Bud6 stimulates the nucleation or elongation phase of filament growth (an important mechanistic distinction), and whether the activities of Bud6 are important in vivo. Graziano and collaborators dissected Bud6 mechanism by: (a) generating mutations in Bud6 that separately disrupt its interactions with formins (bu6-35) and actin monomers (bud6-8), (b) using TIRF (total internal reflection fluorescence) microscopy to visualize the effects of Bud6 and formins on individual actin filaments polymerizing in real time, and (c) performing a genetic analysis of bud6 alleles. They made three important observations. First, Bud6 enhances the nucleation rather than elongation phase of actin assembly, in sharp contrast to another formin ligand, profilin, that enhances elongation. Second, this activity of Bud6 requires its direct interactions with both the formin and actin monomers, suggesting that Bud6 recruits monomers to the formin to help assemble an actin ‘seed’. Third, genetic perturbation of these activities of Bud6 results in reduced levels of actin cable formation in vivo, in turn causing defects in polarized secretion and cell growth.

Until now, formins were thought to nucleate actin assembly by themselves, which is mechanistically distinct from the Arp2/3 complex (another major actin nucleator). Efficient nucleation by Arp2/3 requires the addition of a nucleation-promoting factor (NPF) such as WASp or WAVE, which recruits actin monomers. Graziano et al. reveal that some formins are similar to Arp2/3 in that they too require an NPF for robust nucleation. Their findings also uncover unanticipated mechanistic parallels between the two systems, since in each case nucleation requires both an actin filament end-capping component (formin or Arp2/3) and an actin monomer-recruiting factor (Bud6 or WASp).

How well is this formin-NPF mechanism conserved? Clues to this question have recently emerged from other studies. A paper published last year in The Journal of Cell Biology by the Goode lab, working in collaboration with the labs of Niko Grigorieff (Brandeis) and Gregg Gundersen (Columbia), implicates the human tumor suppressor protein Adenomatous polyposis coli (APC) in functioning as a formin NPF (Okada et al., 2010). Another study published in The Proceedings of the National Academy of Sciences by the labs of Mike Eck (Dana Farber Cancer Institute), Margot Quinlan (UCLA), and Avital Rodal (Brandeis), suggests that Spire, which is conserved in mammals and flies, may serve as a formin NPF (Vizcarra et al., 2011). Bud6, Spire, and APC all bind multiple actin monomers and interact with the C-terminus of formins to enhance actin assembly, suggesting that they may have related mechanisms and perform functionally analogous roles.

Although the requirement of NPFs increases the complexity of the formin mechanism, it offers an explanation for how cells simultaneously overcome two prominent barriers to actin assembly found in vivo – actin monomer binding proteins (e.g. profilin) that suppress formation of an actin nucleus and capping proteins that terminate growth by associating with the growing end of the filament. NPFs can facilitate nucleation by recruiting actin monomers in the presence of profilin, and formins protect growing ends of filaments from capping proteins. Future work will focus on identifying new formin-NFP pairs, defining the cellular processes with which they are associated, and distinguishing the underlying mechanistic differences among each set.

Who pulls the strings in actin cable assembly?

When large structures are built inside of cells, how are their dimensions determined? Are cues received that tell the structure to keep growing, or to slow down, or to stop growing altogether? A recent study published in Developmental Cell by a team led by Molecular and Cell Biology PhD student Melissa Chesarone-Cataldo and Professor of Biology Bruce Goode begins to address these questions by focusing on cytoskeletal structures called yeast actin cables.

Actin cables serve as essential railways for myosin-dependent transport of vesicles, organelles and other cargo, required for yeast cells to grow asymmetrically and produce a daughter cell. Cables are assembled at one end of the mother cell and run the length of the entire cell, but no longer, or else they would hit the back of the cell, buckle and misdirect transport. So how does an actin cable know how long to grow? How are other properties of the cable, such as its thickness and mechanical rigidity determined, and how important are these properties for cable function in vivo?

Actin cables are assembled at the bud neck by the formin protein Bnr1, and rapidly extend into the mother cell at a rate of 0.5-1 µm/s. At this speed, the tip of the actin cable reaches the back end of the cell in about 5-10 seconds. Each cable consists of many shorter overlapping pieces (individual actin filaments) that are stitched or cross-linked together to form a single cable, and cables continuously stream out of the bud neck due to the robust actin assembly activity of Bnr1. Chesarone-Cataldo et al. asked the question, “what mechanism prevents the cables from colliding with the back of the cell and overgrowing?” In doing so, they identified a novel actin cable ‘length sensing’ feedback loop, dependent on the myosin-passenger protein Smy1.

Using live-cell imaging, they showed that Smy1 molecules are transported by myosin from the mother cell to the bud neck, where they pause to interact with the formin Bnr1. Purified Smy1 attenuated Bnr1 activity by slowing down the rate of actin filament elongation. When the SMY1 gene was deleted, cables grew too long, hit the rear of the cell and buckled (see image, right). In addition, the mutant cables abnormally fluctuated in thickness and were kinked, impairing transport of myosin and its cargoes.

The authors propose that a negative feedback loop controls actin cable length. In their model, the cargo (Smy1 in this case) communicates with the machinery that is making the cable (the formin Bnr1), as a means of sensing ‘railway’ length. The longer the railway grows, the more passengers it picks up, and the more transient inhibitory pulses the formin receives. As such, longer cables are selectively attenuated, while shorter cables are allowed to grow rapidly. This negative feedback loop allows yeast cells to tailor actin cable length to the dimensions of the cell and to the needs of its myosin-based transport system.

Current work in the Goode lab is aimed at testing many of the mechanistic predictions of the model above and understanding how Smy1 functions in coordination with other known regulators of Bnr1, all simultaneously present in a cell, to produce actin cables with proper architecture and function. In addition, experiments are underway to find out whether related mechanisms are used to control formins in mammalian cells and to understand the physiological consequences of disrupting those mechanisms.

Chesarone-Cataldo M, Guérin C, Yu JH, Wedlich-Söldner R, Blanchoin L, Goode BL. The Myosin passenger protein Smy1 controls actin cable structure and dynamics by acting as a formin damper. Dev Cell. 2011 Aug 16;21(2):217-30.

A molecular function of Zillion Different Screens protein explained

In a recent paper in Journal of Cell Biology entitled “Spatial regulation of Cdc55-PP2A by Zds1/Zds2 controls mitotic entry and mitotic exit in budding yeast“, Brandeis postdoctoral fellow Valentina Rossio and Assistant Professor of Biology Satoshi Yoshida reveal a molecular function of a mysterious protein Zds1.

The Zds1 protein in yeast  was identified some years ago in “a zillion different screens” for cell cycle mutants, stress response mutants, RNA metabolism mutants, etc., but the molecular function of the protein remained a mystery for more than 15 years. Rossio revealed that Zds1’s key target is a protein phosphatase 2A (PP2A) complex. She showed that Zds1 controls nucleocytoplasmic distribution of PP2A complex, and that this regulation is critical for cells to know when to enter and to exit from mitosis (picture below; cells lacking Zds proteins adopt an abnormal shape because of problems in mitosis). Rossio thinks all the other complicated phenotypes associated with ZDS1 can also be explained by PP2A regulation and is currently studying mechanistic details about the Zds1-PP2A interaction.

See also the accompanying commentary “Proteins keep Cdc55 in its place

Yeast genetics and familial ALS

In a recent paper in PLoS Biology, “A Yeast Model of FUS/TLS-Dependent Cytotoxicity“, Brandeis postdoc Shulin Ju and coworkers applied yeast genetics to examine the function of the human protein FUS/TLS. The gene for FUS/TLS is mutated in 5-10$ of cases of Familial ALS. The yeast model expressing the mutant protein recapitulates many important features of the pathology.

A particular feature of interest is that  FUS/TLS form cytoplasmic inclusions of this protein which is normally localized to the nucleus. Over-expression of a number of yeast proteins rescues the cells from the toxic effect without removing the inclusions. The results are suggested to implicate RNA processing or RNA quality control in the mechanism of toxicity, which I find really interesting in light of the talk Susan Lindquist (an author on this paper) gave at Brandeis about yeast prions and regulatory proteins earlier this month.

Other authors on the paper include Brandeis professors Dagmar Ringe and Gregory Petsko, and Brandeis alumni Dan Tardiff (PhD, Mol. Cell. Biol.,  ’07), currently a postdoc in the Lindquist lab at the Whitehead Institute,  and Daryl Bosco (PhD, Bioorganic Chem, ’03), currently on the faculty at U. Mass. Medical School.

For more information, please see the paper itself or the longer article about the research on Brandeis NOW.

Phosphatases and DNA double strand break repair

When cells suffer DNA damage – as little as a single break in one chromosome – they respond by activating the DNA damage checkpoint, which prevents cells from entering mitosis until there is enough time to to repair the damage.  The principal biochemical events in the checkpoint pathway are the phosphorylations of protein kinases by other protein kinases and eventually the phosphorylation of other proteins that regulate mitosis.    When repair is complete, the checkpoint must be turned off.  Not surprisingly, the enzymes that turn off the checkpoint are phosphatases that can remove the phosphates added by the protein kinases.

The Haber lab has previously shown that, in budding yeast, a pair of PP2C phosphatases known as Ptc2 and Ptc3 were important in turning off a key protein kinase, Rad53.  A member of another phosphatase subgroup, the PP4 phosphatase Pph3, dephosphorylates a target of the checkpoint kinases, histone protein H2A.  There is one aspect that they didn’t understand at all: It seems that the intensity of the checkpoint signals must grow the longer it takes to repair DNA damage, because deletions of ptc2 and ptc3 or a deletion of pph3 prevented cells from turning off the damage signal when it took a long time – 6 hours – to repair the damage, but they had much less effect on different repair events that could complete in 3-4 hours or in less than 2 hours.  So they decided to see what would happen if they created a yeast strain lacking all three phosphatases (ptc2 ptc3 pph3), leading to a paper appearing this month in the journal Molecular and Cell Biology.

To their surprise, these cells had a new defect: they couldn’t complete the repair event itself, rather than simply being defective in resuming mitosis after repair was completed.  The mutants could not properly initiate the small amounts of DNA copying that are required for repair.  Again, the severity of the defect depends on the length of the delay it takes to initiate the repair event itself.  The figure (right) shows that the triple mutant is also much more sensitive to DNA damaging agents such as the anti-cancer drug camptothecin (CPT) and to methylmethansulfonate (MMS). These data show a complex connection between DNA damage signaling and the repair process itself, and reveal new roles for the phosphatases in DNA repair.  The work was carried out primarily by graduate student Jung-Ae Kim, now a postdoc at Rockefeller University, with help by another grad student, Wade Hicks, and by an undergraduate Sue Yen Tay, and postdoc Jin Li. The work was supported by a research and a graduate student training grant from the NIH.

Protected by Akismet
Blog with WordPress

Welcome Guest | Login (Brandeis Members Only)