Towards a P Systems Pseudomonas Quorum Sensing Model

Towa ds a P Sys ems Pseudomonas Quo um Sensing Model Luca Bianco1, Da io Pescini2, Pe e Siepmann3, Na alio K asnogo 3, F ancisco J. Rome o-Campe o4, and Ma ian Gheo ghe5 1Depa men o Compu e Science, Uni e si y o Ve ona S ada Le G azie 15, 37134 Ve ona, I aly [email p o ec ed] 2Dipa imen o di In o ma ica, Sis emis ica e Comunicazione Uni e si `a degli S udi di Milano-Bicocca Via Bicocca degli A cimboldi 8, 20126 Milano, I aly [email p o ec ed] 3School o Compu e Science and In o ma ion Technology Uni e si y o No ingham Jubilee Campus, No ingham, NG81BB, UK {Pe e .Siepmann, Na alio.K asnogo }@no ingham.ac.uk 4Resea ch G oup on Na u al Compu ing Depa men o Compu e Science and A ificial In elligence Uni e si y o Se ille, A da. Reina Me cedes, 41012 Se illa, Spain [email p o ec ed] 5Depa men o Compu e Science, The Uni e si y o Sheffield Regen Cou , Po obello S ee , Sheffield S1 4DP, UK [email p o ec ed] Abs ac . Pseudomonas ae uginosa is an oppo unis ic bac e ium ha exploi s quo um sensing communica ion o synch onize indi iduals in a colony and his leads o an inc ease in he effec i eness o i s i ulence. In his pape we de i ed a mechanis ic P sys ems model o desc ibe he beha io o a single bac e ium and we discuss a possible app oach, based on an e olu iona y algo i hm, o une i s pa ame e s ha will allow a quan i a i e simula ion o he sys em. 1 In oduc ion The quo um sensing is a pa icula o m o cell- o-cell communica ion in bac e ia which exploi s he concen a ion o a pa icula molecule, called signal, o “sense” he popula ion densi y o he colony. The quo um sensing egula o y ne wo k is used by he indi iduals o he colony o collec i e synch oniza ion and he e o e o a cohe en con ol o e he gene exp ession. In Pseudomonas ae uginosa his mechanism is esponsible o he effec i eness o he i ulence o his bac e ium [14,23,10,9]. In ac , a single bac e ium s a s o exp ess his i ulence ac o s only when i senses ha he bac e ia popula ion has eached a ce ain h eshold le el such ha he hos esponse will be inadequa e. The ac i a ion o a complex cellula esponse is wha dis inguishes he quo um sensing as a communica ion egula o y ci cui om o he densi y dependen esponses such as he me aboliza ion o de oxifica ion o small molecules. The simples quo um sensing ne wo k known in G am-Nega i e bac e ia is also he fi s one e e disco e ed [25,17]. I has been ound in he Vib io fische i bac e ium, also known as Pho obac e ium fische i and is nowadays conside ed as he pa adigm o his cell communica ion p ocess. In his ne wo k wo p o eins and one signalling molecule a e in ol ed. The Rp o einis a ansc ip ional egu- la o , while he Ip o einis he syn hase o he signalling molecule, also e e ed o as he au oinduce . An impo an ole is also played by he confinemen o he bac e ial colony. The ac ha he au oinduce molecule is no dispe sed in he en i onmen allows i s diffusion inside he indi iduals and he e o e i s concen a ion sensing. A low cell densi ies he Ip o einsyn hesizes he au oinduce a a basal a e and he signal eely diffuses ou side he bac e ium. The concen a ion o he sig- nal inside each bac e ium is inc eased by he combined effec o he confinemen and he inc ease o he popula ion. A his poin , he binding o he Rp o ein wi h he au oinduce becomes mo e likely. The binding o he signal molecules ac i a es he Rp o ein ansc ip ional egula o . Since he I gene is he a ge o he Rp o ein, he bac e ium s a s o p oduce mo e and mo e signal. The egula ion ne wo k signal au oinduces i s ansc ip ion. In his way he high con- cen a ion o he au oinduce coo dina es he ansc ip ion o all he genes ha a e a ge o he Rp o ein. The quo um sensing in Pseudomonas ae uginosa is mo e complex, ne e heless in iguing, since his bac e ium uses wo diffe en quo um sensing sys ems which in e ac wi h each o he . The aim o his wo k is o p o ide a P sys em model [15,16] o he bac e ium Pseudomonas ae uginosa quo um sensing ocusing on he communica ion mech- anisms. The pa ame e s o he model will be uned using an e olu iona y al- go i hm. Ou long e m aim is o ep oduce he cha ac e is ic beha io o he quo um sensing in Pseudomonas ae uginosa, namely, he swi ch be ween wo dis inc s able s eady solu ions: he fi s desc ibing he beha io o he non- quo a ed bac e ium (i.e., wi h low le els o au oinduce ), he second modeling i s quo a ed beha io (i.e., he beha io ob ained wi h high concen a ion o he au oinduce molecule). Once he model will be en i ely defined se e al simula- ions wi h diffe en s a egies [6,20,18] will be un. Fi s o all, we add ess he modeling o he in e nal dynamics o one single bac e ium, uning i s kine ic cons an s in a way ensu ing i s non-quo a ed be- ha io . A a la e s age, we in end o exploi compa men aliza ion o P sys ems o model a colony o bac e ia each o hem in e nally specified acco ding o he same se o kine ic cons an s. In his espec we will ex end he cu en model o a Popula ion P sys ems app oach [3] ha has been al eady used o exp ess some aspec s o quo um sensing in bac e ium Pseudomonas ae uginosa [22] and o sel -assembly p oblems [4]. 2 An Ini ial Model The fi s s age o ou in es iga ion is in ended o desc ibe he quo um sensing ela ed ne wo k o each bac e ium o cap u e i s main ea u es in o a mecha- nis ic model. The quo um sensing in e nal pa hway o each bac e ium is aken om models discussed in [11,9] and a g aphical ep esen a ion o all elemen s in ol ed in i , as well as some ele an ela ionships be ween hem, a e depic ed in Figu e 1. LasR LasR lasR lasI LasI 3O V saL RsaL Fig. 1. The Pseudomonas quo um sensing model analyzed he e ( om [9]). No e ha double a ows deno e e e sible eac ions, bold ones he deg ada ion p ocess and he emp y ones he inhibi o y p ocess. Acco ding o his model, he quo um sensing pa hway comp ises wo in e con- nec ed signalling cascades. The main elemen s in ol ed in he fi s one a e p o- eins LasR,RsaL,LasI (as well as he genes in ol ed in hei p oduc ion), he au oinduce molecule 3-oxo-C12-HSL and he ac i e complex LasR-3-oxo-C12- HSL. The key elemen s o he second sys em a e he p o eins RhlR and RhlI (as well as he genes in ol ed in hei p oduc ion), he au oinduce molecule C4- HSL and he ac i e complex RhlR-C4-HSL. The fi s one o he wo signalling cascades is called las sys em because i was shown o egula e he exp ession o LasB elas ase. This pa hway egula es o he i ulence ac o s such as LasA p o ease, exo oxin A, alkaline p o ease A as well as he exp ession o a leas wo genes o he xcp sec e o y pa hway. The las pa hway is posi i ely con olled by GacA and V whe eas i is inhibi ed by RsaL ha , in u n, is posi i ely egu- la ed by he ac i e complex LasR-3-oxo-C12-HSL and whose ole is o ep ess he ansc ip ion o he lasI gene. The second signalling sys em in ol ed in he model is named hl sys em be- cause i con ols he exp ession o hamnolipid ia he p oduc ion o hlAB ope on. The au oinduce molecule in his case is C4-HSL and he ac i e com- plex is RhlR-C4-HSL. I has been shown ha his cascade is necessa y o he p oduc ion o some i ulence ac o s like LasB elas ase and LasA p o ease, as well as pyocyanin, cyanide and alkaline p o ease. Fo his eason his signalling sys em is also known as sm ( i ulence seconda y me aboli es). Al hough he co esponding au oinducing molecules a e highly selec i e (and hus no in e changeable a all), se e al in e connec ions be ween he las and he hl pa hways o he quo um sensing in Pseudomonas ae uginosa a e known. One link be ween hem has been al eady men ioned and i is cons i u ed by he LasB elas ase, ha needs bo h LasR-3-oxo-C12-HSL and RhlR-C4-HSL o i s p oduc ion. Mo e in e es ingly, he las sys em is a a highe le el in he hie a chical egula o y cascade, in ac LasR-3-oxo-C12-HSL can ac i a e he exp ession o he hlR gene. In addi ion, he ac i e complex LasR-3-oxo-C12- HSL canbind oRhlR p e en ing i o o m he complex RhlR-C4-HSL. 2.1 The Diffe en ial Equa ion Model Many models o he quo um sensing in he Pseudomonas ae uginosa a e p e- sen ed in li e a u e and usually hey app oach he phenomenon om wo diffe - en angles. The fi s one desc ibes he colony beha io by summa izing indi idual dynamics as a s a e change a oiding a p ecisely de ailed ep esen a ion o each o he bac e ium quo um sensing ne wo ks [24,1]. The second one desc ibes in a mo e de ailed ashion he quo um sensing pa hway o each bac e ium wi h he pu pose o model he eme gen beha io o he whole colony [11]. We hink ha he P sys em amewo k is pa icula ly sui able o his second app oach. In ac , he modula i y, he compa men aliza ion, he hie a chical s uc u e and he ew i ing ules (all ea u es o P sys ems [16]) allow a con e- nien desc ip ion o his eali y. In [11] a model o he las signalling sys em has been de ised, bu no desc ip- ion is gi en o he hl sys em. The g aphical desc ip ion o he quo um sensing pa hway depic ed in Figu e 1 has been ansla ed in o he se o eigh diffe en ial equa ions p esen ed in he nex page. The co espondence be ween diffe en ial equa ion symbols and elemen s in he pa hway a e summa ized in Table 1. The p oduc ion o he ac i a ed complex Pby means o he au oinduce and he LasR p o ein (whose exp ession is gi en by he p oduc o he cons i u i e elemen s concen a ions wi h a a e kRA:kRARA)isanexampleo howcoope - a i e con ibu ions a e ob ained in he diffe en ial equa ions app oach by means o he mass ac ion law. Basal a es p oduc ions and deg ada ions a e also aken in o accoun , an ex- ample o he o me being he k1 elemen gi ing he basal p oduc ion o LasR p o ein (R), while an example o he la e is he deg ada ion o he ac i e com- plex (P) ep esen ed by he elemen kPP. The p oduc ion o messenge RNAs om he co esponding genes is modeled wi h a Michaelis-Men en-like dynam- ics depending on he concen a ion o he p omo ing ac o , as i happens in he case o he p oduc ion o lasR and saL mRNAs ( espec i ely and s), he fi s modeled by V P K +Pand he second by Vs P Ks+P. The p oduc ion o lasI mRNA (l) is also down- egula ed by he p esence o RsaL p o ein (S)and his is modeled by Vl P Kl+P 1 KS+S, in which he Michaelis-Men en-like dynamics is a enua ed by an in e sely p opo ional unc ion o he RsaL concen a ion. dP d =kRARA −kPP dR d =−kRARA +kPP−kRR+k1 dA d =−kRARA +kPP+k2L−kAA dL d =k3l−kLL dS d =k4s−kSS ds d =Vs P Ks+P−kss d d =V P K +P−k + 0 dl d =Vl P Kl+P 1 KS+S−kll+l0 (1) Un o una ely, no alue is known o he 21 kine ic cons an s p esen in he se o diffe en ial equa ions (1). To o e come his p oblem, in [11] se e al simpli ying assump ions a e conside ed, ha lead o ewe equa ions and ewe pa ame e s as well. In he ollowing we will desc ibe a possible pa ame e es ima ion s a egy o ackle his p oblem (see Sec ion 4). The idea is o elay o his diffe en ial equa ions sys em as a “syn he ic bio-expe imen ” used o con on ou model o. 2.2 A Fi s P Sys ems Model Se e al a emp s o simula e he quo um sensing in bac e ia a e p esen in P sys ems li e a u e [5,19], bu , as a as we know, none o hem deals wi h he Pseudomonas ae uginosa bac e ium. He e we desc ibe a di ec P sys ems ansla ion o he diffe en ial equa ion model p e iously discussed [11]. Fo mally, he Pseudomonas P sys em is Π=(A, μ, w, R) Table 1. Va iable-concen a ion co espondence be ween he diffe en ial o mula ion and he g aphical desc ip ion o he quo um sensing model o Pseudomonas ae uginosa ( om [11]) Va iable Concen a ion RLasR A3-oxo-C12-HSL PLasR-3-oxo-C12-HSL LLasI SRsaL lasR mRNA llasI mRNA s saL mRNA whe e: –A={geneR,geneL,R,A,P,L,S, ,l,s}is he alphabe ; –μ=[] 0is he memb ane s uc u e: since we add ess he single bac e ium case, i con ains he cellula memb ane only; –w=geneR geneL is he ini ial configu a ion ha comp ises only LasR and LasI genes, hus is ep esen ed as he s ing; –R={ 1,···, 18}is he se o he ules: 1:geneR −→ geneR + 2: −→ λ 3: −→ +R 4:P−→ P+ 5:R+A−→ P 6:P−→ R+A 7:P−→ P+s 8:s−→ λ 9:S−→ λ 10 :s−→ s+S 11 :P−→ P+l 12 :l−→ l+L 13 :l−→ λ 14 :geneL −→ geneL +l 15 :L−→ λ 16 :L−→ L+A 17 :A−→ λ 18 :R−→ λ No e ha , symbols in Aco espond o he a iables o he diffe en ial equa ion and hei co espondence o he biological eali y is gi en in Table 1. Two new elemen s (i.e., geneR and geneL) a e in oduced, which accoun o he genes in ol ed in he basal p oduc ion o he LasR and LasI mRNAs. Each one o he ules in Ris di ec ly ob ained om he diffe en ial desc ip ion o he conside ed quo um sensing model. Fo examples, we can see ha ule 1 models he basal p oduc ion o he LasR mRNA, while ule 2exp esses i s deg ada ion, mo eo e ules 5and 6desc ibe he e e sible eac ion o he complex P o ma ion by s a ing om i s undamen al cons i uen s Rand A. Due o he diffe en le el o abs ac ion in he ep esen a ion o diffe en pa s o he model (as in he case o he Michaelis-Men en-like kine ics ha a e modelled wi h a highe le el o abs ac ion han o he componen s o he sys em), we canno di ec ly apply mechanis ic algo i hms [2] o his model. Fo his eason, we will apply o his se o ules only he s a egy known as Me abolic Algo i hm ( o de ails e e o [6]), whose simula ion esul s, oge he wi h some nume ical solu ions o he se o diffe en ial equa ions (1), a e shown in Sec ion 2.3 o diffe en choices o pa ame e s. The me abolic algo i hm simula ion needs o speci y a se o eac ion maps, each one associa ed in a one- o-one manne o he ules o R. Reac ion maps [6] a e unc ions defined o e he s a e o he sys em (i.e., mul iplici y o concen- a ion o all elemen s o he sys em depending on he case), ha a e used by he Me abolic algo i hm o alloca e objec s o ules. Fo example, as we will see in a while, F 1, ha is he eac ion map o ule 1, is simply he cons an a e o p oduc ion o LasR mRNA. We can ha e mo e complica ed eac ion maps, as in he case o ule 4 ha akes in o accoun he Michaelis-Men en-like p o- duc ion o he LasR mRNA elici ed by he LasR-3oxo-C12-HSL complex. As in he case o he ules, ha speci y he physical in e ac ions and connec ions be ween he elemen s o he modeled eali y, we can ob ain his in o ma ion om he diffe en ial equa ion o mula ion. The se o eac ion maps employed in ou simula ions a e he ollowing: F 1= 0F 2=k F 3=k1F 4=V K +P F 5=kRA F 6=kP F 7=Vs Ks+PF 8=ks F 9=k4F 10 =kS F 11 =Vl (Kl+P)·(KS+S)F 12 =k3 F 13 =klF 14 =l0 F 15 =kLF 16 =k2 F 17 =kAF 18 =kR (2) No e ha all eac ion maps a e cons an apa om h ee o hem. We ha e al- eady discussed he meaning o he eac ion map associa ed o ule 4; analogous conside a ions hold o F 7as well. Mo e in e es ing is he eac ion map associ- a ed o ule 11 ha akes in o accoun he inhibi o y effec o RsaL p o ein on he p oduc ion o he lasI mRNA. Rema kably, he me hod allows he cu en desc ip ion o diffe en pa s o he sys em a diffe en abs ac ion le els; mo eo e i is s ill applicable i all eac ion maps a e cons an , a condi ion equi ed by mechanis ic algo i hms. In he ollowing some simula ion esul s a e shown, as well as he nume ical solu ion o he diffe en ial equa ion sys em, o some chosen pa ame e s. 2.3 Simula ion Resul s He e we show how he same model- eali y can be desc ibed wi h wo diffe en app oaches. As men ioned be o e, we do no ha e p ecise alues o he model pa ame e s. Fo his eason, as a fi s compa ison a emp , we make a comple ely fic i ious choice o hem. As a u he wo k, we plan o adop some au oma ic way o he pa ame e es ima ion (see Sec ion 4 o mo e de ails). The ini ial choice o pa ame e s is he e shown, and all he subsequen changes o his ini ial pa ame e se will be explici ly men ioned: kRA =10 kP=2 kR=5 k1=1 k2=1 kA=1 k3=1 kL=1 k4=1 kS=1 Vs=1 Ks=1 ks=0.5V =1 K =1 k =1 0=1 Vl=1 Kl=1 kl=1 l0=1 KS=1 (3) The lack o biological in o ma ion makes his choice comple ely a bi a y and p e en s us o compu e he dynamics o he sys em by means o s ochas ic algo i hms such as he Gillespie one [12,13], Dynamical P obabilis ic P Sys ems [20] o he Mul i-compa men al Gillespie [18]. In his sec ion we compa e he dynamics gene a ed by he me abolic algo- i hm wi h he solu ions ob ained o he co esponding diffe en ial equa ion sys em. Figu e 2 depic s he case in which pa ame e s a e chosen acco ding o (3). The dynamics o each species eaches a s eady s a e in bo h app oaches, bu he ela i e posi ion o he species is diffe en and his leads o wo dis inc sys em dynamics. Mo eo e , he ime o he wo sys ems diffe s; in he solu ion o he diffe en ial equa ion sys em his is measu ed in a bi a y uni s (due o he a bi a y choice o pa ame e s), while in he model based on P sys ems he ime is measu ed in s eps o sys em e olu ion. In Figu e 3 he choice o Vl=0 swi ches off ule 11 o he P sys em model and in his case he esul s o he wo diffe en app oaches quali a i ely ma ch each o he . Finally, he las choice o pa ame e s is aimed a ob aining a quo um sensing consis en beha io , ha is, in he case o a single bac e ium in he en i onmen i should no quo a e and hus he concen a ion o he complex Pshould each he basal a e. Ac- co dingly, we se KRA o he alue 0.1. In his case, depic ed in Figu e 4, he dynamics p oduced by he wo app oaches is quali a i ely simila again. 0 0.5 1 1.5 2 2.5 0 5 10 15 20 25 30 P R A L S s l 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 5000 10000 15000 20000 25000 30000 P R A L S s l Fig. 2. Resul s o he quo um sensing model wi h pa ame e s showed in (3) using ODE app oach (le ) and me abolic algo i hm ( igh ) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 5 10 15 20 25 30 P R A L S s l 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 5000 10000 15000 20000 25000 30000 P R A L S s l Fig. 3. Resul s o he quo um sensing model wi h Vl= 0 using ODE app oach (le ) and me abolic algo i hm ( igh ) 0 0.2 0.4 0.6 0.8 1 1.2 0 5 10 15 20 25 30 P R A L S s l 0 0.2 0.4 0.6 0.8 1 1.2 0 5000 10000 15000 20000 25000 30000 P R A L S s l Fig. 4. Resul s o he quo um sensing model wi h pa ame e s kRA =.1usingODE app oach (le ) and me abolic algo i hm ( igh ) 3 Towa ds a De ailed P Sys ems Model Al hough he p elimina y P sys em model desc ibed in Subsec ion 2.2 showed ha we can ob ain compa able esul s wi h he cu en models p esen ed so Fig. 8. The a ge Michaelis-Men en concen a ions and he e ol ed P sys ems ones 5 Conclusions and Fu he Wo k We ha e b iefly desc ibed a pa o he quo um sensing ne wo k in he Pseudomo- nas ae uginosa. S a ing om a diffe en ial equa ions based model we ha e p o ided a P sys ems e sion o i and we compa ed he dynamics o he wo app oaches. In o de o apply diffe en simula ion s a egies on his in iguing phenomenon we p o ided a mo e de ailed, mechanis ic model which, we belie e, is close o he biological eali y. The lack o biological in o ma ion ega ding he dynamics o he sys em led us o use an au oma ic way o es ima ing hem by using an e olu iona y algo i hm app oach ha offe s a eliable and effec i e me hod in his espec . An immedia e s ep u he , a e ob aining all he pa ame e s egula ing a single bac e ium dynamics, is o ex end he p oposed model a a colony le el, exploi ing he compa men aliza ion offe ed by P sys ems and al eady es ab- lished popula ion P sys ems models. O he impo an de elopmen s a e ela ed o he use o expe imen al da a o une he dynamics o ou specifica ions such as o simula e eal biological p ocesses. In his espec he use o model checking me hodologies, al eady unde conside a ion in a pape unde p epa a ion, will con ibu e owa ds alida ing ce ain p ope ies o he sys ems modeled. On long e m we belie e ha hese s eps can ep esen he fi s s age owa d a quan i a i e analysis ha will hope ully lead o a success ul d ug design p ocess. Acknowledgemen s. N. K asnogo and P. Siepmann acknowledge he EPSRC o unding p ojec EP/D021847/1. Re e ences 1. K. Anguige, J.R. King, J.P. Wa d, and P. Williams. Ma hema ical modelling o he apies a ge ed a bac e ial quo um sensing. Ma hema ical Biosciences, 192:39– 83, 2004. 2. A. P. A kin. Syn he ic cell biology. Cu en Opinion in Bio echnology, 12:638–644, 2001. 3. F. Be na dini and M. Gheo ghe. Popula ion P sys ems. Jou nal o Uni e sal Compu e Science, 10:509–539, 2004. 4. F. Be na dini, M. Gheo ghe, N. K asnogo , and J.-L. Gia i o. On sel -assembly in popula ion P sys ems. In C.S. Calude, M.J. Dinneen, G. Pˇaun, and M.J. Pe´ ez- Jime´nez, edi o s, Uncon en ional Compu a ion. 4 h In e na ional Con e ence, UC 2005, pages 46–57, 2005. 5. F. Be na dini, M. Gheo ghe, N. K asnogo , R.C. Muniyandi, M.J. P´e ez-Jim´enez, and F.J. Rome o-Campe o. On P Sys ems as a modelling ool o biological sys- ems. In R. F eund, G. Lojka, M. Oswald, and Gh. Pa˘un, edi o s, P e-P oceedings o he 6Th In e na ional Wo kshop on Memb ane Compu ing (WMC6), pages 193– 213, 2005. 6. L. Bianco, F. Fon ana, and V. Manca. P Sys ems wi h Reac ion Maps. In e na- ional Jou nal o Founda ions o Compu e Science, 17(1):27–48, 2006. 7. G.E. B iggs and J.B.S. Haldane. A no e on he kine ics o enzyme ac ion. Biochem. J., 19:338–339, 1925. 8. K. A. Conno s. Chemical Kine ics: The s udy o Reac ion Ra es in Solu ion.VCH, 1990. 9. C.V. Delen and B.H. Iglewski. Cell- o-cell signalling and Pseudomonas ae uginosa in ec ions. Eme ging In ec ious Diseases, 4(4):551–560, Oc obe -Decembe 1998. 10. S.P. Diggle, K. Winze , A. Lazdunski, P. Williams, and M. C´ama a. Ad anc- ing he quo um in Pseudomonas ae uginosa: M aT and he egula ion o N- acylhomose ine lac one p oduc ion and i ulence gene exp ession. Jou nal o Bac- e iology, 184:2576–2586, 2002. 11. J.D. Docke y and J.P. Keene . A ma hema ical model o quo um qensing in Pseudomonas ae uginosa.Bulle in o Ma hema ical Biology, 63:95–116, 2001. 12. D.T. Gillespie. A gene al me hod o nume ically simula ing he s ochas ic ime e olu ion o coupled chemical eac ions. Jou nal o Compu a ional Physics, 22:403– 434, 1976. 13. D.T. Gillespie. Exac s ochas ic simula ion o coupled chemical eac ions. Jou nal o Compu a ional Physics, 81(25):2340–2361, 1977. 14. A.M. Lazdunski, I. Ven e, and J.N. S u gis. Regula o y ci cui s and communica- ion in G am-nega i e bac e ia. Na u e Re iews, Mic obiology, 2:581–592, 2004. 15. G. P˘aun. Compu ing wi h memb anes. J. Compu . Sys em Sci., 61(1):108–143, 2000. 16. G. P˘aun. Memb ane Compu ing. An In oduc ion. Sp inge , Be lin, 2002. 17. J.P. Pea son. Ea ly ac i a ion o quo um sensing. Jou nal o Bac e iology, 184:2569–2571, 2002. 18. M.J. Pe´ ez-Jime´nez and F. J. Rome o-Campe o. P sys ems – A new compu a- ional modelling ool o sys ems biology. T ansac ions in Compu a ional Sys ems Biology, 2006 (in p ess). 19. M.J. P´e ez-Jim´enez and F.J. Rome o-Campe o. Modelling Vib io fische i’s be- ha iou using P sys ems. In Sys ems Biology Wo kshop, ECAL, 2005. 20. D. Pescini, D. Besozzi, G. Mau i, and C. Zand on. Dynamical p obabilis ic P sys ems. In e na ional Jou nal o Founda ions o Compu e Science, 17(1):183, 2006. 21. P.A. Siepman, G. Te azas, and N. K asnogo . E olu iona y Design o he Be- ha iou o Cellula Au oma on-Based Complex Sys ems. In P oceedings o he Se en h In e na ional Con e ence on Adap ing Compu ing in Design and Manu- ac u e. 22. G. Te azas, N. K asnogo , M. Gheo ghe, F. Be na dini, S. Diggle, and M. Cama a. An en i onmen awa e P sys em model o quo um sensing. In S. Ba y Coope , B. L¨owe, and L. To en lie , edi o s, New Compu a ional Pa adigms. Fi s Con . on Compu abili y in Eu ope, CiE2005, pages 479–485, 2005. 23. A.U. Vi e a and M. Fussenegge . Modelling he quo um sensing egula o y ne - wo k o human-pa hogenic Pseudomonas ae uginosa.Bio echol. P og., 20:670–678, 2004. 24. J.P. Wa d, J.R. King, A.J. Koe be , P. Williams, J.M. C o , and R.E. Socke . Ma hema ical modelling o quo um sensing in bac e ia. Jou nal o Ma hema ics Applied in Medicine and Biology, 18:263–292, 2001. 25. K. Winze , K.R. Ha die, and P. Williams. Bac e ial cell- o-cell communica ion: so y, can’ alk now – gone o lunch! Cu en OpinoninMic obiology, 5:216–222, 2002.