From gradients to stripes: a logical analysis of the genetic network controlling early drosophila segmentation.

Transcription

1 From gradients to stripes: a logical analysis of the genetic network controlling early drosophila segmentation. Denis THIEFFRY (LGPDIBDM, Marseille, France) Contents 1. Logical modelling of gene networks 2. Application to Drosophila segmentation 3. Conclusions and prospects

2 Logical formalisation (> R. Thomas et al.) Fragment of a regulatory graph Gene I Gene J Gene R Multilevel logical variables and functions Here, r, R takes their values in {0, 1} or {0, 1, 2} Gene Q Occam razor : minimisation of the number of distinct logical values

3 Logical parameters Gene I Gene J Gene R + parameters > effects of combinations of interactions!!! 1 parameter > effect of each individual interaction 1 parameter > basal expression Gene Q All these parameters > take the same value as the corresponding variable/function > may present inter dependencies (inequalities) Default situation ( Occam razor ): logical parameters > 0

4 Drosophila Segmentation Collaboration with Lucas SANCHEZ (CIB, Madrid) Source: Wolpert et al. (1998)

5 Anteriorposterior patterning system Collection and integration of regulatory data (GINdb) Graph analysis Three strongly connected components: Gap Pairrule Segmentpolarity crossregulatory modules Source: Wolpert et al. (1998)

6 Models of the Gap Module gt hb Kr kni gt + hb + () Kr + kni () gt hb Kr kni gt 2 1/3 2 0 hb Kr kni Reinitz (1996) Bodnar (1997) gt hb Kr kni gt 0 0 hb Kr +/ 0 kni + 0 gt hb Kr kni gt 0 0 hb 0 (+) 0 Kr +/ 0 kni 0 0 RiveraPomar & Jäckle (1996) Sánchez & Thieffry (2001)

7 Gap Module Bcd Hb mat Cad Maternal Zygotic gap Gt Hb zyg Kr Kni

8 Multilevel logical model for the Gap module gap network maternal inputs gt hb Kr kni bcd cad hb mat gt hb 0 (+1) 2 0 +[1...3] 0 (+1) Kr 1 +1/ kni No restriction on multiple interactions Thresholds considered as parameters Sensitivity to context (K s) 7 feedback circuits: 2 positive circuits: gtkr, gtknikrhb 1 conditional autocatalytic circuit: hb 3 negative circuits: gtkrhb, gtknikr, hbknikr 1 dual (+/) circuit: hbkr

9 Gap module: logical parameter values Gene (A) b=3, c=0, h m =2 (B) b=2, c=0, h m =2 (C) b=1, c=1, h m =0 (D) b=0, c=2, h m =0 giant K g.b 0 K g.b 0 K g.b 0 K g.c 0 K g.bh 0 K g.bh 0 K g.bh 0 K g.ch 0 K g.br 1 K g.br 1 K g.br 1 K g.cr 0 K g.bhr 1 K g.bhr 1 K g.bhr 1 K g.chr 1 K h.ε 3 K h.δ 2 K h.b 1 K h 0 hunchback Krüppel knirps K h.εr 3 K h.δr 2 K h.ρ 1 K h.h 0 K h br 1 K h.r 0 K h.ρr 1 K h.hr 0 K r.b 0 K r.b 0 K r.b 0 K r. 0 K r.bg 0 K r.bg 0 K r.bg 0 K r.g 0 K r.bh 0 K r.bh 0 K r.bh 0 K r.h 0 K r.bn 0 K r.bn 0 K r.bn 0 K r.n 0 K r.bgh 1 K r.bgh 1 K r.bgh 1 K r.gh 0 K r.bgn 0 K r.bgn 0 K r.bgn 0 K r.gn 0 K r.bhn 1 K r.bhn 1 K r.bhn 1 K r.hn 0 K r.bghn 2 K r.bghn 2 K r.bghn 2 K r.ghn 0 K n.b 0 K n.b 0 K n.bc 0 K n.c 0 K n.bg 0 K n.bg 0 K n.bcg 0 K n.cg 0 K n.bh 0 K n.bh 0 K n.bch 0 K n.ch 0 K n.bgh 0 K n.bgh 0 K n.bcgh 1 K n.cgh 1 Anterior pole Posterior pole

10 Gap Module Simulation (gt, hb zyg, Kr, kni) Bcd=3, hb mat =2, cad=0 Bcd=2, hb mat =2, cad=0 Bcd=1, hb mat =0, cad=1 Bcd=0, hb mat =0, cad=2 [1300] gt hb Kr kni [0220] [0111] [1000] hb bcd Kr kni gt cad gt

11 Gap Module Simulation (gt, hb, Kr, kni) Head Trunk Telson BCD HB mat BCD HB mat BCD CAD CAD Initial conditions: overlapping maternal gradients multiple (asynchronous) transitions BCD HB GT BCD HB KR BCD CAD HB KR CAD GT Four patterns of gap gene expression KNI

12 Simulation of maternal and gap lossoffunction mutations Genetic background Final state (GT, HB, KR, KNI) A B C D Wildtype Bicoid Observations/predictions loss of GT in region A loss of HB in ABC and of KR in BC KNI expands anteriorly into region AB Hunchback mat No significative effect increase of KR in region C caudal loss of KNI in region C loss of GT in region D giant KNI expands posteriorly into D Krüppel GT expands into regions B and C Loss of KNI in region C knirps increase of KR in region C Hunchback mat&zyg GT expands into regions B and C loss of KR in regions B and C loss of KNI in region C giantkrüppel KNI expands posteriorly into region D Krüppelknirps GT expands into regions B and C giantknirps increase of KR in region C Anterior pole 4 trunk domains Posterior pole

13 Simulation of gap gene ectopic expression Genetic background Final state (GT, HB, KR, KNI) A B C D Wildtype giant Hunchbacklevel 1 Hunchbacklevel 2 Hunchbacklevel 3 Krüppellevel 1 Krüppellevel [12] [12] Observations/predictions lowering of KR in region B loss of KNI in region C. loss of GT and activation of KNI in D loss of GT in region D activation of GT in BD but not in D loss of GT in AD, KNI activated in D loss of GT in AD, KNI activated in D knirps loss of KR and activation of GT in B giant + knirps Loss of KR in region B and C Anterior pole 4 trunk domains Posterior pole

14 Simulation of cisregulatory mutations Final state Genetic (GT, HB, KR, KNI) background A B C D Wildtype Hb autoregulation 0[12]20 0[12] Kr BS controlling Gt Observations/predictions activation of KR and repression of GT in region A expansion of GT in regions B and C, > further repression of KR and KNI in regions B and C, respectively Anterior pole 4 trunk domains Posterior pole

15 Pairrule genes expression patterns Oddnumbered parasegment Evennumbered parasegment Oddnumbered parasegment Hairy Run Hairy Eve Eve Run Early expression Ftz Odd Ftz Odd Ftz Odd Prd Prd Prd Slp Slp Slp Run Run Run Eve Eve Eve Odd Odd Odd Late expression Ftz Prd Prd Prd Prd Slp Slp Slp Region I Region II Region III Region Region IV V

16 Pairrule crossregulations Oddnumbered parasegment Evennumbered parasegment Oddnumbered parasegment Eve Ftz Eve Run Odd Run Prd Prd Slp Slp Region I Region II Region IIIa + IIIb Region IV Region V

17 Pairrule crossregulatory gene network wg slp odd opa ftz eve ppa run prd en

18 Ouputs (Segment polarity genes) Pairrule crossregulatory matrix with thresholds eve prd ppa run slp ftz odd opa eve prd ppa run slp ftz odd en wg Input (ubiquist expression)

19 Gene value 1 value 2 value 3 eve K v.vrs K v.psd K v.vsd K v.prsd K v.vpsd K v.vprs K v.vrsd K v.vprsd prd ppa run slp K p.d K p.vd K p.zd K p.zvd K a.v K r.pvd K s.vzd K p.ad K p.vad K p.azd K p.zvad Logical parameter values ftz K z.vsd K z.zvsd odd en wg K d.vp K d.zvp K e.yzs K e.yprsd K e.yzrsd K e.ypzsd K e.ypzrsd K w.ypsvzd

20 Intrinsic dynamical properties of the pairrule module V P A R S Z D [ ] [ ] [ ] [ ]

21 Pairrule system and positional information The state reached by the pairrule system depends on which of the pairrule genes become first activated (3 rules) depends on the position along the A/P axis of the embryo s trunk depends on particular combination of maternal & gap products along the A/P axis of the embryo s trunk

22 Logical simulations for the wild type Region I Region II Region IIIa Region IIIb Region IV Region V V P A R S Z D V P A R S Z D V P A R S Z D V P A R S Z D V P A R S Z D V P A R S Z D Activation of primary pairrule genes C1 C C C C1 C C1 C Activation of secondary pairrule genes and refinement of pairrule stripes activation of prd, ppa and slp activation of prd, ppa and slp activation of prd, ppa and slp activation of prd, ppa and slp [ ] [ ] activation of prd, ppa and slp Final state [ ] [ ] [ ] [ ] [ ] wg = ON & en = OFF wg = OFF & en = ON wg = OFF & en = OFF wg = ON & en = OFF wg = OFF & en = ON odd parasegment even parasegment odd parasegment

23 Three rules for the formation of the pairrule pattern C1) Eve odd overcomes Odd eve C2) maternalgap inputs > run overcomes Eve run C3) Hairy ftz & Hairy run are needed

24 Simulation of lossoffunction pairrule mutants Genetic background wildtype eve stable states v p a r s z d prd ppa run slp ftz odd Embryo regions III I, IV II V IV Not reached Not reached I, IIIV II III I, IV II V III IV I, II,V III I, II, (IV) (IV), V IIV V I, IIIb, IV II, IIIa V EN/WG expression () () (partially) functional circuits eve (+), eve/run (+), eve/slp (+), prd/odd (+), slp/ftz (+), eveftzslp (), prd/odd/ftz (+) prd/odd (+), slp/ftz (+), prd/odd/ftz (+) eve (+), slp/ftz (+) eve (+), eve/run (+), eve/slp (+), prd/odd (+), slp/ftz (+), eveftzslp (), prd/odd/ftz (+) eve (+), eve/slp (+), prd/odd (+), slp/ftz (+), eveftzslp (), prd/odd/ftz (+) eve (+), eve/run (+), prd/odd (+), prd/odd/ftz (+) eve (+), eve/run (+), eve/slp (+), prd/odd (+) eve (+), eve/run (+), eve/slp (+), slp/ftz (+), eveftzslp () Comments Loss of all and stripes Loss of oddnumbered stripes and all stripes and stripes are normally formed (though three of the wildtype stable states are modified). Loss of oddnumbered stripes Replacement of stripes by stripes Loss of evennumbered stripes and expansion of stripes Evennumbered and stripes expand posteriorly and anteriorly into the parasegment, respectively

25 Prediction of double lossoffunction mutants Genetic stable states background wildtype eve;odd ftz;odd prd;slp Embryo regions III I, IV II V I, II, IIIa IIIb, IV, V IIV V III I, II, IV, V EN/WG expression (partially) functional circuits eve (+), eve/run (+), eve/slp (+), prd/odd (+), slp/ftz (+), eveftzslp (), prd/odd/ftz (+) slp/ftz (+) eve (+), eve/run (+), eve/slp (+) eve (+) Comments Odd stripes are replaced by stripes; even stripes expand posteriorly; odd stripes are replaced by stripes; even stripes expands anteriorly. Even stripes are lost and the stripes are expanded. Loss of oddnumbered stripes and of all stripes

26 Simulation of ectopic expression of pairrule genes Genetic stable states background v p a r s z d wildtype Embryo regions III I, IV II V EN/WG expression eve IV None! prd ppa run slp ftz odd I, IIIb,IV II, IIIa V I, IV II III V III I, IV, V II III I, II, IV, V III I, II, IV V IV Not reached () (partially) functional circuits eve (+), eve/run (+), eve/slp (+), prd/odd (+), slp/ftz (+), eveftzslp (), prd/odd/ftz (+) eve (+), eve/run (+), eve/slp (+), slp/ftz (+), eveftzslp () eve (+), eve/run (+), eve/slp (+), prd/odd (+), slp/ftz (+), eveftzslp (), prd/odd/ftz (+) eve (+), prd/odd (+), slp/ftz (+), prd/odd/ftz (+) prd/odd (+) eve (+), eve/run (+), prd/odd (+) eve (+) Comments The stripes are lost and the stripes are expanded Anterior expansion of stripes Loss of odd stripes Replacement of odd stripes by stripes Replacement of stripes by stripes Replacement of stripes by stripes Loss of and stripes

27 Prediction of pairrule cisregulatory mutants Genetic stable states background v p a r s z d wildtype eve autoregulation Ftz autoregulation Embryo regions III I, IV II V III I, IV, (V) II, (V) III I, IV II V EN/WG expression (partially) functional circuits eve (+), eve/run (+), eve/slp (+), prd/odd (+), slp/ftz (+), eveftzslp (), prd/odd/ftz (+) prd/odd (+), slp/ftz (+), prd/odd/ftz (+) eve (+), eve/run (+), eve/slp (+), prd/odd (+), slp/ftz (+), eveftzslp (), prd/odd/ftz (+) Comments Possible replacement of odd stripes by stripes Loss of even stripes

28 Feedback circuit analysis 51 intertwined, distinct feedback circuits but only 7 (partly) functional feedback circuits: 6 +, 1 Multiple positive circuits on a given gene (e.g. eve) cooperation to generate a nonlinear switch!? Perturbations at the feedback circuits Loss of certain stable states (cellular states) Misslocalization of some of the remaining states

29 Four expression modes 1. The eveexpression mode formation of oddnumbered stripes 2. The ftzexpression mode formation of evennumbered stripes 3. The prd/slp expression mode formation of odd and evennumbered stripes 4. The oddexpression mode prevention of en and wg expression in the middle of the parasegment

30 Segment polarity gene network slp fz zw3 arm dsh wg en Ci Ci[act] Ci[rep] fu pka Ptchh hh smo ptc ptc, hh, Ci[act] are represented by quaternary variables; Ptchh, wg, smo, pka, fz, dsh are represented by ternary variables; the others 8 regulatory products are represented by binary variables.

31 Segment polarity intercellular network fz fz slp slp zw3 arm zw3 arm wg dsh en wg dsh en Ci Ci Ci[act] Ci[rep] fu pka hh fu pka hh PtcHh smo ptc smo ptc Anterior cell Parasegmental border Posterior cell

32 Summing up Emphasis on the roles of regulatory circuits (+ and ) vs gene cascades Crucial role of positive circuits in differentiative decisions (Krgt) Principally short circuits are found functional (Krgt, eve, everun...) Notion of "crossregulatory modules" = sets of intertwined circuits Notion of "crossregulatory modes" Crossregulatory circuits and modules as dynamical building blocks Qualitative reproduction of the gene expression patterns associated with the different cell types Prediction of the phenotypes of various types of mutants

33 Prospects Coupling between the gap, pairrule and segment polarity modules towards a model of the whole segmentation hierarchy Modelling of the network controlling cell cycle analysis of the coupling between cell cycle and cell differentiation Comparative and evolutionary analysis of homologous regulatory networks (graph topology, qualitative dynamics, redundancy) Methodological/computational developments (formalism, software) From Logical Regulatory Graphs to Petri Nets

34 Multidisciplinary collaborations at Luminy LGPDIDBDM: Anaïs BAUDOT Christine BRUN Claudine CHAOUYIA Aitor GONZÁLEZ Carl HERRMANN Bernard JACQ David MARTIN Pierre MOUREN Denis THIEFFRY IML: Alain GENOCHE Badih GHATTAS Brigitte MOSSE Elisabeth REMY LIF: Victor CHEPOI Claude SABATIER Yann VAXES Michel Van Caneghem CPT: Pierre CHIAPPETTA Bastien FERNANDEZ Pierre GIRAUD Sébastien JAEGER André LAMBERT Ricardo LIMA Arnaud MEYRONEINC