Cell cycle regulation by feed‐forward loops coupling transcription and phosphorylation

• DOI: 10.1038/msb.2008.73
• Publication Date: 2009-01-20T13:24:34Z
• Authors: Attila Csikász‐Nagy; Orsolya Kapuy; Attila Tóth; Csaba Pál; Lars Juhl Jensen; Frank Uhlmann; John J Tyson; Béla Novák
• Keywords: 0301 basic medicine; 570; Medicine (General); Transcription, Genetic; QH301-705.5; Cell Cycle; 500; DNA replication; feed‐forward loop; Cyclin-Dependent Kinases; 03 medical and health sciences; R5-920; Report; budding yeast; cell cycle; Biology (General); Phosphorylation; Transcription Factors

Report20 January 2009Open Access Cell cycle regulation by feed-forward loops coupling transcription and phosphorylation Attila Csikász-Nagy Corresponding Author The Microsoft Research––University of Trento Centre for Computational Systems Biology, Povo (Trento), Italy Search more papers this author Orsolya Kapuy Department Biochemistry, Oxford Integrative University Oxford, UK Applied Biotechnology Food Science, Budapest Technology Economics, Budapest, Hungary Tóth Csaba Pál Biology Unit, Institute Biological Research Center, Szeged, Lars Juhl Jensen European Molecular Laboratory, Heidelberg, Germany Novo Nordisk Foundation Center Protein Research, Copenhagen, Denmark Frank Uhlmann Chromosome Segregation Cancer London Institute, London, John J Tyson Sciences Virginia Bioinformatics Polytechnic & State University, Blacksburg, VA, USA Béla Novák Information 1, Kapuy2,3, Tóth3, Pál1,4, Jensen5,6, Uhlmann7, Tyson8 2,3 1The 2Department 3Department 4Systems 5European 6Novo 7Chromosome 8Department *Corresponding authors. Research—University Piazza Manci 17, 38100, Italy. Tel.: [phone]; Fax: 2814; E-mail: [email protected] South Parks Road, OX1 3QU, UK. [phone]; 613213; (2009)5:236https://doi.org/10.1038/msb.2008.73 PDFDownload PDF article text main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures Info eukaryotic cell requires precise temporal coordination the activities hundreds 'executor' proteins (EPs) involved in growth division. Cyclin-dependent protein kinases (Cdks) play central roles regulating production, activation, inactivation destruction these EPs. From genome-scale data sets budding yeast, we identify 126 EPs that are regulated Cdk1 both through direct EP factors control expression EP, so each is a loop (FFL) from Cdk1. By mathematical modelling, show such FFLs can activate at different phases depending effective signs (+ or −) regulatory steps FFL. We provide several case studies controlled exactly as our models predict. signal-transduction properties allow one (or few) Cdk signal(s) drive host responses correct sequence. Introduction A cell's progression G1, S, G2 M replication division orchestrated large-amplitude fluctuations kinase (Cdk) generated series coupled positive negative feedback (Novak et al, 2007; Holt 2008; Skotheim Novak, 2008). signals transduced into appropriate specific executor (Sutani 1999; Tanaka 2007a) (Box 1). For example, components signalling pathway called 'mitotic exit network' yeast 'septation initiation fission (Bardin Amon, 2001). Recently, showed (Csikasz-Nagy 2007) septation network has characteristic topology (FFL): high level Cdk1–cyclin B mitosis activates function early (sensors) inactivates late (executors). High activity primes network, but cannot 'fire' until falls releases inhibitory arm. similar FFL controls onset DNA synthesis, according 'licensing factor' hypothesis (Blow, 1993). Recognizing executing synthesis division, hypothesized might be common motifs transmitting master complexes target execute events. Box 1 signal transduction Feed-forward loops, involving substrate (EP) its factor (TF), proposed transducers between (periodic activity) responses, substrates potential EPs, periodically expressed during (Spellman 1998; 2006). Intriguingly, often (Ubersax 2003; Furthermore, (TFs) cycle-dependent gene must cycle-regulated themselves, it reasonable suspect least some them phosphorylated Cdks. Wherever case, Cdk–TF–EP trio an Owing large-scale experimental screens (Saccharomyces cerevisiae) targets Loog Morgan, 2005), well TFs (Lee 2002), possible systematically test genome-wide scale. Results discussion To end, classified all 4691 verified protein-coding genes genome 6 non-overlapping topologies (Figure 1A) based on whether not encoded been reported substrate, known TF target. identified FFL, Of FFLs, 68 (54%) found cycle, whereas only 13 would expected chance (P<10−28). None other shows comparably ratio 1A; Supplementary Table S1). Thus, clear strong predictor periodicity involvement motif. This observation suggests transcribed, FFL-regulated (Supplementary S2) may indeed key Figure 1.FFL-regulated over-represented among transcribed cycle-related genes. (A) All ORFs were distributed groups their (Cdk1) factors. group, report number transcribed/total proteins. details, see S1. (B) Odds ratios (observed/expected) finding with certain type (as explained (A)) MIPS functional category term given colour code legend. detailed statistics, S3. On six panels, single star denotes those cases where probability random appearance (according binomial distribution) less than 10−3, two stars 10−6. dashed line indicates odds 1. Download figure PowerPoint further support assertion, functions significantly FFLs. checked distribution (and related) annotations (Ruepp 2004; Guldener 2005) established 1A. most classes 1B; S3). converse statement also true: terms associated S4). conclude important 'signals' 'responses' over-representation small group 'only Cdk'-regulated If conclusion correct, then, once discovered, will fall disproportionately group. FFL-regulatory topologies, then ask what function(s) pathways orchestrating cycle. depends three links motif (± ±/±). first sign activation − inhibition) effect Cdk-mediated TF, second active form upregulates downregulates expression. product net (activation 'long arm' activity. third inhibits it. eight combinations divided (Mangan Alon, 2003): coherent (±±/+) (∓±/−) same long short arms incoherent (±±/−) (∓±/+) opposite signs. Coherent have noise-filtering when sustained (in S+G2+M phase), absent prolonged period time G1 phase). Incoherent rich response capabilities (Tyson Csikasz-Nagy Soyer, Kaplan particular relevance here, they respond sufficiently bursts signal: activated transiently rises after low (at G1/S transition), M/G1 transition). propose many bioinformatics survey genome/proteome how regulate events, study dynamics theoretical perspective. model using ordinary differential equations reactions delay changes concentrations 2A; S5). implement transient per arm lower threshold operate faster timescale indirect These differences arise naturally phosphorylation-transcription FFL: happens within seconds, delayed production (timescale∼minutes) (Adelman 2002). 2.Four limit attention here upregulation TF; downregulation, information. Four types EP. Arrows + represent inhibition, respectively. Computer simulations S5) describing interactions diagrammed above. Black line: activity; coloured lines: (A). Proposed borders indicated. Simulation results shown 2B. In figure, plot black) typical trajectory think 'signal generator' transducers' begins rise transition, peaks rapidly cells return phase. As expected, (− +/−) +/+), phase (yellow curve) (red curve), transition (blue curve: +/+) FFL) (green FFL). ensure proper G1-specific (S+G2+M)-specific convert periodic strict alternation S-phase entry M-phase exit, transitions occur duplication separation genetic material. Next, use diverse evidences predict, cases, effects S6). predictions, could 59 46 which S7). examples including regulators whose times match predictions theory 2B). 3, controlling protein, Sic1 (Knapp 1996), mitotic Cdc5 (Zhu 2000), Dbf2 (Visintin 2001) initiator, Sld2 (Tanaka 2007a). Sld2, database search revealed Cdk' (with expression). However, Ash1 (Teixeira 2006) Sld2. should predicted activator SLD2 prediction fits recent role 2007b; Zegerman Diffley, fluctuation profile (not shown) (Masumoto 3.Examples transcriptional post-translational controls. Interaction (±) rules presented S6. Both polo (Cdc5) (Fkh2) presumably (bound B-type cyclins). initiator shares Cdc5, appears inhibited PEST sequence (Rechsteiner Rogers, 1996) phosphoprotein detected (Chi 2007), suggesting induces degradation. (C) inducer 2007). Although there no documented Our predicts (D) stabilizer, Sic1, directly Swi5 1996). (E) An example complex embedding Further details S7. basic described theoretically clearly oversimplifications schemes operating real cells. 3C) illustrates overlapping even contradictory. contains sequences S6), that, (Zegerman 2007b), site degradation, giving overlapping, contradictory Similar initiators replication, MCM Cdc6. (Our methods insufficient early, before degraded.) Cln3 3E) interlocked employed achieve effects. 3D) presents composed double-negative loop, because well-known inhibitor Cdk1-Clb (Schwob 1994). (=positive) switch, flipping (Cdk1-Clb high) start off low) (Chen 2004). switch made robust. feature demonstrated recently removing sites (Cross i.e. leg passing, note Cdk1-Cln, Cln-dependent do simple 3A) multiple downstream targets, Cdc25, Wee1 cyclin activating (Barr Activation turns pair stabilizing operational homologues Cdc25 entry. continually either (when high). Consulting 1A, 'chain' serve purposes equally well. But robust single-arm 2003). robustness advantage: needed induce events (DNA division) order. analysis playing 1B) predict inducers under 3B C). Altogether, suggest converting oscillations topologies. Conclusion organisms studied detail, appear Cdk–cyclin pairs crucial coordinating Each own suite probably Nonetheless, sufficient viable (Fisher Nurse, simulation 2B) shows, principle, pair, utilizing four motifs, G1- G2-specific trigger alternating manner. imagine last ancestor present-day relied signal, played oscillatory coordinated eukaryotic-style idealized view 1) provides scenario evolution eukaryotes merit now 'first approximation' organization present day organisms, exploit modify intrinsic dynamical potentials. still intimately yeast. Materials obtained 2005). downloaded YEASTRACT act complexes, say if total, 600 de Lichtenberg al (2005). FunCat CYGD (Guldener information, determining TF–EP connections Cdk1-mediated phosphorylations. Model construction wrote S4) rates change forms only. assume thus degraded, total amount represents form. Parameters chosen get unique course was minimal system, comparable Acknowledgements thank Fredrick R Cross stimulating discussions Orkun S Soyer critical reading paper. supported grants Hungarian Scientific Fund (OTKA-F60414), Italian Ministry Project FIRB (RBPR0523C3) (AC-N), Council (202591) (CP), Commission (YSBN FP7: 201142), National Institutes Health (5R01GM079207) James McDonnell (21002050) (BN JJT). Conflict Interest authors declare conflict interest. Supporting S1 (application/jpg, 642.4 KB) S2 2.8 MB) S3 2.2 S4 2.4 (application/xls, 369 60 105.5 1,009 S5 (PDF document, 14.7 S6 61.5 S7 40 (application/doc, 90 References Adelman K, La Porta A, Santangelo TJ, Lis JT, Roberts JW, Wang MD (2002) Single molecule RNA polymerase elongation reveals uniform kinetic behavior. Proc Natl Acad Sci 99: 13538–13543CrossrefCASPubMedWeb Science®Google Scholar Bardin AJ, Amon (2001) MEN SIN: what's difference? Nat Rev Mol Biol 2: 815–826CrossrefCASPubMedWeb Barr FA, Sillje HH, Nigg EA (2004) Polo-like orchestration 5: 429–440CrossrefCASPubMedWeb Blow JJ (1993) Preventing re-replication cycle: evidence licensing factor. 122: 993–1002CrossrefCASPubMedWeb Chen KC, Calzone L, FR, Novak B, 15: 3841–3862CrossrefCASPubMedWeb Chi Huttenhower C, Geer LY, Coon JJ, Syka JE, Bai DL, Shabanowitz J, Burke DJ, Troyanskaya OG, Hunt DF (2007) Analysis Saccharomyces cerevisiae electron transfer dissociation (ETD) mass spectrometry. 104: 2193–2198CrossrefCASPubMedWeb Schroeder Bean JM Phosphorylation cyclins essential contributes robustness. Genetics 176: 1541–1555CrossrefCASPubMedWeb O, Gyorffy Modeling (SIN) Curr Genet 51: 245–255CrossrefCASPubMedWeb OS (2008) Adaptive two-state protein. Soc Interface 5 (Suppl 1): S41–S47CrossrefCASPubMedWeb U, LJ, Brunak Bork P (2005) Dynamic formation Science 307: 724–727CrossrefCASPubMedWeb Fisher Nurse (1996) p34cdc2 promotes absence cyclins. EMBO 850–860Wiley Online LibraryCASPubMedWeb Munsterkotter M, Kastenmuller G, Strack N, van Helden Lemer Richelles Wodak SJ, Garcia-Martinez Perez-Ortin Michael H, Kaps Talla E, Dujon Andre Souciet JL, De Montigny Bon Gaillardin Mewes HW CYGD: Comprehensive Yeast Genome Database. Nucleic Acids Res 33: D364–D368CrossrefCASPubMedWeb Krutchinsky AN, Morgan DO Positive sharpens anaphase switch. Nature 454: 353–357CrossrefCASPubMedWeb TS, (2006) Co-evolution cell-cycle regulation. 443: 594–597CrossrefCASPubMedWeb Bren Dekel Alon U generate non-monotonic input Syst 4: 203Wiley LibraryPubMedWeb Knapp D, Bhoite Stillman Nasmyth K regulates p40Sic1. 16: 5701–5707CrossrefCASPubMedWeb Lee TI, Rinaldi NJ, Robert F, Odom DT, Bar-Joseph Z, Gerber GK, Hannett NM, Harbison CT, Thompson CM, Simon I, Zeitlinger Jennings EG, Murray HL, Gordon DB, Ren Wyrick Tagne JB, Volkert TL, Fraenkel Gifford DK Transcriptional networks cerevisiae. 298: 799–804CrossrefCASPubMedWeb Cyclin specificity cyclin-dependent substrates. 434: 104–108CrossrefCASPubMedWeb Mangan (2003) Structure 100: 11980–11985CrossrefCASPubMedWeb Zaslaver feedforward serves sign-sensitive element networks. 334: 197–204CrossrefCASPubMedWeb Masumoto Muramatsu Kamimura Y, Araki H S-Cdk-dependent chromosomal 415: 651–655CrossrefCASPubMedWeb Irreversible due systems-level feedback. 9: 724–728CrossrefCASPubMedWeb Rechsteiner Rogers SW proteolysis. Trends Biochem 21: 267–271CrossrefCASPubMedWeb Ruepp Zollner Maier Albermann Hani Mokrejs Tetko Mannhaupt FunCat, annotation scheme systematic classification whole genomes. 32: 5539–5545CrossrefCASPubMedWeb Schwob Bohm T, Mendenhall MD, (1994) p40sic1 S. 79: 233–244CrossrefCASPubMedWeb JM, Di Talia Siggia ED, FR ensures 291–296CrossrefCASPubMedWeb Spellman PT, Sherlock Zhang MQ, Iyer VR, Anders Eisen MB, Brown PO, Botstein Futcher (1998) identification microarray hybridization. 3273–3297CrossrefCASPubMedWeb Sutani Yuasa Tomonaga Dohmae Takio Yanagida (1999) Fission condensin complex: non-SMC subunits condensation Cdc2 Cut3/SMC4. Genes Dev 13: 2271–2283CrossrefCASPubMedWeb Tak YS, (2007a) CDK step eukaryotes. Div 16CrossrefCASPubMedWeb Umemori Hirai (2007b) CDK-dependent Sld3 initiates 445: 328–332CrossrefCASPubMedWeb Teixeira MC, Monteiro P, Jain Tenreiro Fernandes AR, Mira NP, Alenquer Freitas AT, Oliveira AL, Sa-Correia I database: tool associations 34: D446–D451CrossrefCASPubMedWeb Sniffers, buzzers, toggles, blinkers: signaling cell. Opin 221–231CrossrefCASPubMedWeb Regulation molecular antagonism, hysteresis irreversible transitions. Theor 210: 249–263CrossrefCASPubMedWeb Temporal 18: R759–R768CrossrefCASPubMedWeb Ubersax JA, Woodbury EL, Quang PN, Paraz Blethrow JD, Shah Shokat KM, Targets 425: 859–864CrossrefCASPubMedWeb Visintin R, Cdc15 Dbf2. 12: 2961–2974CrossrefCASPubMedWeb Diffley JF 281–285CrossrefCASPubMedWeb Zhu Volpe Davis TN, (2000) Two forkhead pseudohyphal growth. 406: 90–94CrossrefCASPubMedWeb Previous ArticleNext Article Volume 5Issue 11 2009In issue FiguresReferencesRelatedDetailsLoading ...