Cellular Stress Responses and Cancer - Semantic Scholar

11 downloads 0 Views 107KB Size Report
activates the checkpoint signalling kinase Rad53. ... Key words: cell cycle, protein phosphorylation, protein serine/threonine kinase. Abbreviations used: ATM, ataxia telangiectasia mutated; BRCA1, breast cancer associated 1;. BRCT, BRCA1 ...
Cellular Stress Responses and Cancer Irish Area Section Meeting

Cellular Stress Responses and Cancer Irish Area Section Meeting held at the National University of Ireland, Galway, on 4–5 July 2002. Edited by M. P. Carty and N. Lowndes (Galway).

Role of the Saccharomyces cerevisiae Rad9 protein in sensing and responding to DNA damage G.W.-L. Toh and N.F. Lowndes1 Genome Stability Laboratory, Department of Biochemistry, National University of Ireland, University Road, Galway, Ireland

Abstract Eukaryotic cells have evolved surveillance mechanisms, known as DNA-damage checkpoints, that sense and respond to genome damage. DNA-damage checkpoint pathways ensure co-ordinated cellular responses to DNA damage, including cell cycle delays and activation of repair mechanisms. RAD9, from Saccharomyces cerevisiae, was the first damage checkpoint gene to be identified, although its biochemical function remained unknown until recently. This review examines briefly work that provides significant insight into how Rad9 activates the checkpoint signalling kinase Rad53.

Introduction Damage to genomic DNA is a significant and constant problem for eukaryotic cells. Environmental mutagens, ionizing radiation and even oxidative metabolism within the cell, can all damage DNA [1]. DNA damage triggers a wide range of cellular responses, including altered gene expression, cell-cycle arrest and stimulation of DNA repair (reviewed in [2]). If left unrepaired, base damage or modification can lead to mutations, while other types of lesions such as strandbreaks and cross-links interfere with processes such as DNA replication, gene transcription and chromosome segregation. The ability to deal with spontaneous or environmentally induced DNA damage is therefore as critical to genome stability and cell viability as accurate DNA replication or proper chromosomal segregation. Consequently, eukaryotic cells have evolved surveillance mechanisms, referred to as DNA-damage checkpoints, that monitor genome integrity and co-ordinate various aspects of the damage response, including cell-cycle progression, gene transcription and DNA repair. A DNA-damage checkpoint is a signal transduction pathway that senses the presence of DNA damage and transmits a signal to downstream effectors, which execute the various cellular responses to DNA damage. Damage checkpoints attenuate the mutagenicity, genomic instability and cell lethality that can result from DNA damage, by delaying cell cycle progression, increasing the transcription

Key words: cell cycle, protein phosphorylation, protein serine/threonine kinase. Abbreviations used: ATM, ataxia telangiectasia mutated; BRCA1, breast cancer associated 1; BRCT, BRCA1 C-terminus; FHA, forkhead-associated; PIKK, phosphoinositide 3-kinase-like kinase. 1 To whom correspondence should be addressed (e-mail [email protected]).

242

 C 2003

Biochemical Society

of genes involved in DNA repair, as well as directly modulating the activity of DNA repair enzymes. In mammalian cells, several damage checkpoint pathway components also function as tumour suppressors, and defects in checkpoint signalling pathways are frequently associated with cancer (reviewed in [2]).

DNA damage checkpoints in budding yeast At present, our most advanced understanding of DNA damage-dependent checkpoints is in the model systems of budding and fission yeast, Saccharomyces cerevisiae and Schizosaccharomyces pombe respectively. Checkpoint proteins are well conserved from yeast to human cells, indicating that the basic organization of these pathways has been preserved throughout evolution. After DNA damage is sensed, an evolutionarily conserved protein kinase cascade amplifies and transmits the checkpoint signal to the various effector pathways. The first level of this kinase cascade consists of members of a family of phosphoinositide 3-kinase-like kinases (PIKKs), which includes Mec1 and Tel1 from budding yeast, as well as mammalian ATM (ataxia telangiectasia mutated), ATR (ATM and Rad3related), and DNA-PKcs (DNA-dependent protein kinase catalytic subunit) (reviewed in [3,4]). These PIKKs control the activation of two classes of unrelated checkpoint signalling kinases, which are represented by the mammalian Chk1 (S. cerevisiae Chk1) and Chk2 (S. cerevisiae Rad53) kinases [5–8]. DNA-damage checkpoints respond to various forms of DNA damage throughout the cell cycle. In budding yeast, DNA damage results in transient cell-cycle arrest in the G1 and/or G2 phases, as well as a slowing of S-phase [9–11].

Cellular Stress Responses and Cancer

Genetic control of these cell cycle delays was first demonstrated by Weinert and Hartwell [11], who showed that the RAD9 gene is required for G2 /M arrest after ionizing irradiation. RAD9 has also been shown to function in the G1 /S [10,12] and intra-S checkpoints [13]. Subsequent genetic analyses identified a further six genes, RAD53, MEC1, RAD17, RAD24, MEC3 and DDC1, that are required for the G1 /S, intra-S and G2 /M checkpoints [14,15].

Rad9: the prototypical damage checkpoint protein Nearly 15 years ago, a pioneering study of budding yeast identified the prototypical DNA-damage checkpoint gene, RAD9 [11]. RAD9 is required for the DNA damage checkpoint in all phases of the cell cycle [9–11], and loss of this gene impairs checkpoint-induced cell-cycle arrest and increases genomic instability [11,16]. Rad9 is a 1309 amino acid protein, with a predicted molecular mass of 148.4 kDa. Its C-terminal region shows localized sequence identity with the mammalian tumour suppressor protein BRCA1 (breast cancer associated 1) [17,18], which contains a tandem repeat of the BRCT (BRCA1 Cterminus) motif (Figure 1). This motif has been shown to be required for Rad9 function, and it mediates the salt-resistant oligomerization of Rad9 after DNA damage [19].

Rad9 hyperphosphorylation links Rad53 to the damage checkpoint Genetic analysis has led to Rad9 being categorized as a sensor/transducer of the damage signal, rather than an effector. Indeed, its biochemical function in activating Rad53 was demonstrated recently by Gilbert et al. [20], who showed that immunoprecipitated hyperphosphorylated Rad9 facilitates the activation of endogenous Rad53 in vitro. It has

also been hypothesized that the protein plays a direct role in sensing and/or processing DNA damage [21], though such a role has so far not been directly demonstrated. Rad9 is required for activation of both Rad53 and Chk1 in response to damage [6,22]. DNA damage, but not blocks to replication, induces Mec1/Tel1-dependent hyperphosphorylation of Rad9, and this hyperphosphorylated form of Rad9 interacts specifically with Rad53 in vivo [23–25]. The Rad9 protein contains 12 potential target sites for phosphorylation by Cdc28, the major cyclin-dependent kinase of budding yeast, as well as 14 potential PIKK phosphorylation sites, six of which lie within the S/T cluster domain of the protein. Rad9 is phosphorylated during normal cell-cycle progression [25], and hyperphosphorylated after DNA damage in an Mec1- and/or Tel1-dependent manner [24,25]. This hyperphosphorylation leads to the association of Rad9 with the checkpoint kinase Rad53, an association that is mediated by interactions between the two forkheadassociated (FHA) domains of Rad53, and phosphorylated residues found specifically on hyperphosphorylated Rad9 [26]. The FHA domains of Rad53 and other Chk2 orthologues are required for their role in the DNA-damage checkpoints. FHA domains are conserved modular protein domains that bind specific phosphopeptides (reviewed in [27]). Rad53 contains two FHA domains, both of which interact with phosphorylated Rad9 [26]. Mutation of conserved amino acids in the second FHA domain (FHA2) disrupts Rad53 activation in response to damage [23]. The [S/T]Q sequence is a target site for phosphorylation by PIKKs such as Mec1, Tel1 and ATM [28]. Using sitedirected mutagenesis, Schwartz et al. [29] identified multiple PIKK target sites within Rad9 that are phosphorylated in response to damage, and that are required for Rad9-dependent activation of Rad53 [29]. Rad9 [S/T]Q phosphorylation site mutants are partially deficient in G2 /M checkpoint function, which requires both Rad53 and Chk1. These mutations do not, however, abolish either the phosphorylation or

Figure 1 Domain structure of Rad9 S. cerevisiae Rad9 is a 1309-amino-acid protein with a predicted molecular mass of 148.4 kDa. It contains two BRCT domains at its C-terminus, which are required for Rad9 oligomerization in response to damage. Sites that are potential targets for phosphorylation by cyclin-dependent kinases (Cdks) and PIKKs (see text) are indicated. NLS, nuclear localization signal; PI3K, phosphoinositide 3-kinase; SCD, S/T cluster domain.

 C 2003

Biochemical Society

243

244

Biochemical Society Transactions (2003) Volume 31, part 1

function of Chk1, implying that they specifically affect Rad53 regulation. Western blot analysis with phospho-specific antibodies directly implicates specific Rad9 [S/T]Q sites as physiological targets for damage-induced phosphorylation, and the relevant Rad9 [S/T]Q phospho-peptides are bound by the FHA domains of Rad53 in vitro. In contrast with multiple Rad9 [S/T]Q mutations, single mutations had no significant effect on the extent of Rad9 hyperphosphorylation, as measured by mobility shift, or on the interaction of phosphorylated Rad9 with Rad53. Individual sites within the Rad9 S/T cluster domain were found to support partial levels of Rad9 phosphorylation and Rad53 interaction, but no single site was capable of fully sustaining either function. This suggests that multiple Rad9 [S/T]Q sites are together responsible for mediating the Rad9–Rad53 interaction [29]. This is consistent with a model in which adsorption of multiple Rad53 molecules occurs, via their FHA domains, during Rad9-mediated catalysis of Rad53 activation [20]. Rad53 is potentially bivalent for interactions with Rad9, since both its FHA domains are able to bind Rad9 phosphopeptides in vitro. The interaction between the two proteins might therefore be stabilized by multivalent contacts between multiply phosphorylated Rad9 and the two Rad53 FHA domains. Furthermore, both Rad9 [19] and Rad53 [29] are capable of homotypic interactions. Thus, all of these interactions may collectively contribute to aggregation of these proteins in checkpoint complexes. It may be that accumulation of such complexes, coupled with the ability of Rad53 (its fission yeast and mammalian homologue Chk2) to selfassociate and activate, promotes a positive-feedback loop that permits massive amplification of the damage checkpoint signal [20,30,31].

Rad9: a biochemical role in Rad53 activation Despite considerable information on the biological regulation of Rad9, until very recently, its biochemical function remained undescribed. Gilbert et al. [20] purified and characterized two distinct Rad9 complexes from undamaged and UVirradiated cells: a large  850-kDa complex containing hypophosphorylated Rad9, and a smaller 560-kDa complex containing hyperphosphorylated Rad9 and also Rad53, which appears after DNA damage. Hydrodynamic analyses indicate that both complexes are of extended elliptical shape, with the larger complex being considerably more extended. Mec1/Tel1-dependent hyperphosphorylation of Rad9 (e.g. after UV irradiation) correlates not only with recruitment of the checkpoint kinase Rad53 [23], but also with the larger complex losing mass and undergoing conformational change. This remodelling of the large Rad9 complex results in the formation of the smaller, less elongated, 560-kDa complex (Figure 2). It is proposed that the latter complex recruits and catalyses the activation of Rad53, by acting as a scaffold that brings Rad53 molecules into close proximity, facilitating in trans autophosphorylation of Rad53. Activated  C 2003

Biochemical Society

Rad53 is subsequently released from the complex in a manner dependent on Rad53 (but not Mec1 or Tel1) kinase activity. No evidence was found for the co-purification of Mec1 with soluble Rad9, indicating that Rad9 does not function by bringing Mec1 into proximity with Rad53. This is consistent with an indirect role for the PIKKs in Rad53 activation, namely that Mec1/Tel1-dependent phosphorylation of Rad9 provides docking sites for Rad53. Confirmation of this has come from the identification of PIKK phosphorylation sites in Rad9, whose mutation disrupts Rad9-dependent regulation of Rad53 [29]. After UV irradiation, the non-phosphorylated forms of Rad53 were found to predominantly co-purify with the hyperphosphorylated Rad9 complex. A clue as to how Rad9 facilitates Rad53 activation comes from the observation that recombinant Rad53 expressed in Escherichia coli becomes rapidly hyperphosphorylated. The extent of this modification is much greater than that seen in yeast cells, and correlates with increasing concentrations of recombinant Rad53. At sufficiently high concentrations in E. coli cells, Rad53 appears to be capable of efficient in trans autophosphorylation, in the absence of other yeast kinases and of Rad9. It has also been shown that efficient in trans autophosphorylation of Rad53 is required for its release from the Rad9 complex. Upon incubation with ATP, kinase-defective, mutant Rad53 is phosphorylated inefficiently, and ATPdependent release of this Rad53 from the 560-kDa Rad9 complex is not observed. Thus, release of phosphorylated Rad53 from the 560-kDa Rad9 complex is an active process requiring hydrolysis of ATP by Rad53. Based on these observations, Gilbert et al. have proposed that the hyperphosphorylated Rad9 complex acts analogously to a solid-state catalyst, where reactants are adsorbed onto a surface, thereby facilitating their reaction to yield products with a lower affinity for that surface, which are then released, allowing further cycles of catalysis to occur. In contrast, no biochemical activity was detected for the larger, hypophosphorylated, Rad9 complex. Neither this, nor the hyperphosphorylated complex, showed appreciable DNA-binding activity to a range of DNA structures (C.S. Gilbert and N.F. Lowndes, unpublished work), which is surprising since Rad9 is predicted to be involved in DNA damage sensing or processing ([32] and references therein). This prediction recently received additional support from studies on green fluorescent protein-tagged, budding yeast checkpoint proteins. Rad9, along with a number of other checkpoint proteins, localizes to nuclear foci following phleomycin treatment (I. Dobbie, personal communication), cdc 13-1 induced telomeric damage, and HO endonuclease induced double strand breaks [33]. HO (HO mothallic switching) endonuclease regulates mating-type switching in budding yeast by introducing a double strand break at the mating-type locus. One possible explanation for the apparent lack of DNA binding activity is that the larger Rad9 complex might simply be a latent form of Rad9 that has to be activated by Mec1 hyperphosphorylation. Another

Cellular Stress Responses and Cancer

Figure 2 Biochemical model for Rad9 function DNA damage induces Mec1/Tel1 phosphorylation of the larger ( 850 kDa) Rad9 complex, which results in its remodelling to the hyperphosphorylated (560 kDa) form. This remodelling involves breakage of salt-sensitive Rad9–Rad9 interactions, and formation of salt-resistant interactions dependent on the Rad9 BRCT domain [19]. The activity of chaperone proteins Ssa1 and/or Ssa2 (Stress Seventy subfamily A 1/2 respectively) is also thought to be required (C.S. Gilbert and N.F. Lowndes unpublished work). Docking of hypophosphorylated Rad53, via its FHA domains, onto the Mec1-generated phosphoresidues in hyperphosphorylated Rad9 increases the local concentration of Rad53 to a level that permits its in trans autophosphorylation. Efficient phosphorylation of Rad53 permits its release from the hyperphosphorylated Rad9 complex, and this activated Rad53 specifically targets substrates required for checkpoint pathway-regulated cell-cycle arrest, gene expression and efficient DNA repair. Ddc2, DNA damage checkpoint 2.

possibility (reviewed in [34]) is that Rad9 interacts indirectly with DNA lesions via so-called ‘adaptors’, for example DNArepair proteins, which would obviate the need for direct contact with DNA.

Implications for other systems Analysis of the primary sequence of Rad9 has so far failed to identify obvious homologues in other systems. In fact, the tandem BRCT repeat at the C-terminus of Rad9 is the only region with appreciable sequence identity with other proteins. This motif is present in many proteins with roles in DNA metabolism [18,35]. Proteins whose BRCT domains appear most related to that in Rad9 include fission yeast Crb2sp /Rhp9sp [36,37], the human tumour suppressor BRCA1, and a p53-binding protein, 53BP1. These proteins share a number of regulatory similarities (reviewed in [34]) that are consistent with the possibility that they may be functional analogues. They are hyperphosphorylated after DNA damage in a manner dependent on the Rad3/Mec1/ ATM class of PIKKs [24,25,36,38], and are also subject to cyclin-dependent kinase-dependent phosphorylation in

S/G2 /M [25,39,40]. In addition, recent evidence indicates that BRCA1 has functions in checkpoint regulation [41,42]. BRCA1 has also been shown to be a component of BASC, a large complex that is speculated to have roles in DNA damage sensing via its interactions with DNA-repair proteins [43]. The structural, regulatory, and biological similarities between these proteins are intriguing, although determination of their biochemical function will be required to adequately address whether these proteins are indeed functional equivalents.

References 1 2 3 4 5 6 7 8 9 10

Lindahl, T. (1993) Nature (London) 362, 709–715 Zhou, B.B. and Elledge, S.J. (2000) Nature (London) 408, 433–439 Abraham, R.T. (2001) Genes Dev. 15, 2177–2196 Durocher, D. and Jackson, S.P. (2001) Curr. Opin. Cell Biol. 13, 225–231 Rhind, N. and Russell, P. (2000) J. Cell Sci. 113, 3889–3896 Sanchez, Y., Bachant, J., Wang, H., Hu, F., Liu, D., Tetzlaff, M. and Elledge, S.J. (1999) Science 286, 1166–1171 Sanchez, Y., Desany, B.A., Jones, W.J., Liu, Q., Wang, B. and Elledge, S.J. (1996) Science 271, 357–360 Sun, Z., Fay, D.S., Marini, F., Foiani, M. and Stern, D.F. (1996) Genes Dev. 10, 395–406 Paulovich, A.G. and Hartwell, L.H. (1995) Cell 82, 841–847 Siede, W., Friedberg, A.S. and Friedberg, E.C. (1993) Proc. Natl. Acad. Sci. U.S.A. 90, 7985–7989  C 2003

Biochemical Society

245

246

Biochemical Society Transactions (2003) Volume 31, part 1

11 Weinert, T.A. and Hartwell, L.H. (1988) Science 241, 317–322 12 Siede, W., Friedberg, A.S., Dianova, I. and Friedberg, E.C. (1994) Genetics 138, 271–281 13 Paulovich, A.G., Margulies, R.U., Garvik, B.M. and Hartwell, L.H. (1997) Genetics 145, 45–62 14 Elledge, S.J. (1996) Science 274, 1664–1672 15 Longhese, M.P., Paciotti, V., Fraschini, R., Zaccarini, R., Plevani, P. and Lucchini, G. (1997) EMBO J. 16, 5216–5226 16 Weinert, T.A. and Hartwell, L.H. (1990) Mol. Cell. Biol. 10, 6554–6564 17 Koonin, E.V., Altschul, S.F. and Bork, P. (1996) Nat. Genet. 13, 266–268 18 Bork, P., Hofmann, K., Bucher, P., Neuwald, A.F., Altschul, S.F. and Koonin, E.V. (1997) FASEB J. 11, 68–76 19 Soulier, J. and Lowndes, N.F. (1999) Curr. Biol. 9, 551–554 20 Gilbert, C.S., Green, C.M. and Lowndes, N.F. (2001) Mol. Cell 8, 129–136 21 Lydall, D. and Weinert, T. (1995) Science 270, 1488–1491 22 Navas, T.A., Sanchez, Y. and Elledge, S.J. (1996) Genes Dev. 10, 2632–2643 23 Sun, Z., Hsiao, J., Fay, D.S. and Stern, D.F. (1998) Science 281, 272–274 24 Emili, A. (1998) Mol. Cell 2, 183–189 25 Vialard, J.E., Gilbert, C.S., Green, C.M. and Lowndes, N.F. (1998) EMBO J. 17, 5679–5688 26 Durocher, D., Henckel, J., Fersht, A.R. and Jackson, S.P. (1999) Mol. Cell 4, 387–394 27 Durocher, D. and Jackson, S.P. (2002) FEBS Lett. 513, 58–66 28 Kim, S.T., Lim, D.S., Canman, C.E. and Kastan, M.B. (1999) J. Biol. Chem. 274, 37538–37543

 C 2003

Biochemical Society

29 Schwartz, M.F., Duong, J.K., Sun, Z., Morrow, J.S., Pradhan, D. and Stern, D.F. (2002) Mol. Cell 9, 1055–1065 30 Ahn, J.Y., Li, X., Davis, H.L. and Canman, C.E. (2002) J. Biol. Chem. 277, 19389–19395 31 Xu, X., Tsvetkov, L.M. and Stern, D.F. (2002) Mol. Cell. Biol. 22, 4419–4432 32 Weinert, T.A. (1998) Curr. Opin. Genet. Dev. 8, 185–193 33 Melo, J.A., Cohen, J. and Toczyski, D.P. (2001) Genes Dev. 15, 2809–2821 34 Lowndes, N.F. and Murguia, J.R. (2000) Curr. Opin. Genet. Dev. 10, 17–25 35 Callebaut, I. and Mornon, J.P. (1997) FEBS Lett. 400, 25–30 36 Saka, Y., Esashi, F., Matsusaka, T., Mochida, S. and Yanagida, M. (1997) Genes Dev. 11, 3387–3400 37 Willson, J., Wilson, S., Warr, N. and Watts, F.Z. (1997) Nucleic Acids Res. 25, 2138–2146 38 Cortez, D., Wang, Y., Qin, J. and Elledge, S.J. (1999) Science, 1162–1166 39 Esashi, F. and Yanagida, M. (1999) Mol. Cell 4, 167–174 40 Ruffner, H., Jiang, W., Craig, A.G., Hunter, T. and Verma, I.M. (1999) Mol. Cell. Biol. 19, 4843–4854 41 Xu, X., Weaver, Z., Linke, S.P., Li, C., Gotay, J., Wang, X.W., Harris, C.C., Ried, T. and Deng, C.X. (1999) Mol. Cell 3, 389–395 42 Larson, J.S., Tonkinson, J.L. and Lai, M.T. (1997) Cancer Res. 57, 3351–3355 43 Wang, Y., Cortez, D., Yazdi, P., Neff, N., Elledge, S.J. and Qin, J. (2000) Genes Dev. 14, 927–939

Received 10 October 2002