Phenotypic variations in a monoclonal bacterial population Oleg Krichevsky, Itzhak Fishov, Dina Raveh, Ben-Gurion University, Beer-Sheva J. Wong, D. Chatenay,

Phenotypic variations in a monoclonal bacterial population

Oleg Krichevsky, Itzhak Fishov, Dina Raveh,Ben-Gurion University, Beer-Sheva

J. Wong, D. Chatenay, M. Poirier, S. Ghozzi, J. Robert

Laboratoire Jean Perrin, FRE 3132 CNRS-UPMC24 rue Lhomond, 75005 Paris

Escherichia Coli Bacteria

1 colony in phase contrast microscopy

4 µm

Electronic microscopy

Schematic bacterium

Cytoplasme (H2O+ions monovalents et divalents)

• Acides nucléiques (ADN, ARN)• protéines (enzymes dont polymérases)• small molecules(ribosome)

Membrane (glycolipide)

•Small numbers of molecules (par ex. 1 chromosome, 10-10000 ARNs; protéines).•Dynamic enzymatic reaction: production, transformation, degradation of the species with time.

Bacterium Biochemistry (simplified!)

2) DNA replicationADNpolymérase, gyrase…

transcription translationADN chromosome ARNm Protéine: un gène

1) Central dogma:

ARNpolymérase ribosome

Growth by division: 1 bacterium→2 daughter bacteria genetically identical (clone)Duplication, repartition of the constituants (in particular of the chromosome)

Division time: 30’à 37°C in nutritive medium(pH~7, protéines, glucides)

Bacterial culture

t

0.01

0.1

1

0 50 100 150 200

Den

sité

opt

ique

(60

0 nm

)

temps(min)

Population/individual• Culture of a single colony in homogenous medium, obtain a monoclonal population

(typically: 1ml de medium grown 12 hours~108 bacteria).

• J. Spudich et D. Koshland revealed the individual character of chemotactism. (Nature 262 1976)

• Mutations don’t explain this individuality→ non geneticorigin.

(mutation rate: 10-10/pb/génération)

• The authors invoked fluctuations of the small number of particle, of chemical rates to explain those non genetic variability.

• This process is more efficient than mutation to allow species adaptation to rapidly fluctutating environnment.

Genetic expression network:

• ADNARNProtéine (fluorescente)

ARN

Gène

PromoteurADN

Taux de transcription kR

Taux de traduction kP

Protéine

Dégradation R

Dégradation P

Example with a negative feedback loop:

• Fluctuations. Network noise. Variability.• Ozbudak et al.: origin of the protein noise expression:

transcription/translation Nature genetics 31 (2002).• Elowitz et al.: Intrinsic noise(Fluctuations des éléments du

réseau)/extrinsic noise(fluctuations des autres composants de la cellule) Science 297 (2002).

• Influence of the regulation mechanism

DNA in bacterium

1 chromosome (4 Mpb)N (1<n<300) plasmid copy number (entre 2 kpb et 100 kpb)

Plasmid

• extrachromosomal DNA fragment

• Code for its copy number (replicon sequence: ori, regulation)

• Uses the host to replicate

• Adds an advantage against otherwise toxic medium (Antibiotic resistance.)

• Symbiotic plasmid/bacterium association

Partition system

Without partition system With partition system

-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

0 20 40 60 80 100 120

Nom

bre

de

cellu

les

Nombre de Copie de Plasmide

Plasmid copy number (PCN) inE. Coli

PCN=phenotype choiceMeasured individual PCN on population scale(~104 individus)

Distribution : variability Antibiotic resistance: adaptability

Standard deviation

<n>

G. Scott Gordon, Dmitry Sitnikov, Chris D. WebbOgden Aurelio Teleman, Aaron Straight, Richard Losick, and Schaechter, Schaech Andrew W. Murray, and Andrew Wright, Cell 1997.

Direct Visualisation directe of plasmids in bacteria:

Fluorescent protein bounds to the plasmid sequence

Disadvantage: homologous recombinaison

Indirect Method

Fluorescent protein mOrange coded by the plasmid[protein] plasmid copy numberFluorescent intensity bacteriaPCN

•Expression copie unique sur le chromosome protéine verte.•Fluorescent gene expression under IPTG inducible tac promoter.

Promoter choice

RNA-polymerase

Promoter tac fluorescent gene Termination seq

LacI repressor→no transcription, no gene expression

IPTG LacI repressor titration→transcription, gene expression

Strong induced promoter: minimise expression noise (Elowitz et al.)

Promoter fluorescent gene Termination seq

RNA-polymerase

Phase contrast Fluorescent image

Measure the fluorescent intensity

Measurement over a population~104, every individual at the same developpment

Low level fluorescence→Flow cytometry+fluorescent microscopy set up

Set up:

cell

optic

detection

Soft lithography microchannel

Mask

photosensitive resin

glassdevelop, fix

Spread PDMS, bake at 90° C

Unmold, fix on a cover glass

UV exposure

Ready to use channel

Optical differentiel Interferometry profilometry image of the channel(z=2µm)

Field of view: 10µmBacterial speed: ~0.1-1 mm/s

Optical elements detail

Time series of fluorescent intensity

FV = PV + AV FO = aPO + AO + PV

Fi: i channel measure of fluorescent intensityPi: i protein fluorescent intensityAi: i channel autofluorescent intensitya: normalization constant between green and orange fluorescence: leak of green toward orange channel

Rq: a posteriori, no orange to green

FO

FV

Bacteria preparation(E. Coli TOP10 strain)

1. Culture 37°C 12h of a clone picked on a petri dish2. Dilution 500X, re culture→DO=0.23. Re dilution 100X, re culture →DO=0.24. Induction 1h 1mM IPTG →protein fluo. production5. Bloque chloramphénicol →stop protein production6. Wash phosphate buffer, 12h. Protein maturation

Bacteria in exponential phase →reproductibilityNo protein production

Limit autofluorescence

Fluorescence level

Calibration

1. "Green" bacteria no plasmid

Induced: leak gren→orange, Non induced :autofluorescence

3. Fluorescent gene in 1 et 2 copies on a plasmid

Linearity between gene copy number and protein expression

2. "Orange" bacteria, no plasmid (=0)

Coefficient a=0.58

Study as a function of the replicon(ampicillin resistance)

• F: single copy, partition system

• R1: low copy number

• ColE1: medium copy number, no partition system

R1+:partition system

R1-:without partition system

F R1- R1+ ColE1

<PV> (a.u.) 27,1 28,5 26,5 25,7

<PO> (a.u.) 27,0 244 173 2167

<n>=<PO>/<PV>=<nP>/<nC> 1,0 7,8 6,5 95

qPCR 0,5 3,2 3,8 23,4

≈constant

Mean plasmid copy number per chromosome

We take <nC>=1,7 (E. Coli and Salomonella, p.1553, ASM Press, 1996)

Hypothesis: On average gene expression does not depend on the copy nor its origin

Variance et variability

F R1- R1+ ColE1

<nP> 1,71 13,3 11 161

0,7 4,2 3,1 40

(%) 46 34 29,2 25 =/<nP>

CPVC

PO

VO

CPP nnP

n

nP

PP

nnn

222

Hypothesis on correlation and autoforrelation of fluorescent protein expression [ <PaPb>, (a,b=O,V)]

0.1

1

10

100

1 10 100 1000

<nP>

Poisson

F

R1's

ColE1

R1- plasmid lossBacteria are cultivated without antibiotic for many generations

with without, 99 générationswithout, 54 générations

Diminution de la Population N+ with plasmid diminishes

Population N- without plasmid increases

Loss rate

• We measure N+(54)=60%, N+(99)=16%

• We deduce: population + division time est higher than 2 min. compared to population –

• Loss rate/bacterium/generation: 0,5%

T

Tcte g

gNgN ,2 )1(

)()(

)1(

1

21

21

)(

)(2

ggN

gN

Boe et Rassmussen, plasmid, 36,p.153 (1996)

10 réactions biochimiques Rµ:

* X0 -> X1 R0 : free promoter -> RNAP bound promoter * X1 -> X0 R1 : unbinding of RNAP freeing promoter * X1 -> X2 + X0 R2 : transcription initiation * X2 -> X3 R3 : transcription, X3 = mRNA * X3 -> Ø R4 : mRNA deggradation * X3 -> X4 R5 : reversible mRNA/ribosome complex formation * X4 -> X3 R6 : reversible mRNA/ribosome complex dissociation * X4 -> X5 + X3 R7 : Translation start freeing RBS * X5 -> X6 R8 : production of protein X6 * X6 -> Ø R9 : protein degradation

Siggia et al., PNAS October 1, 2002 vol. 99 no. 20 12795

transcription traductionADN chromosome ARNm ProteinARNpolymérase ribosome

Numerical simulations

We have M reactions Rµ (m=1,2,…,M) involving N species.We define P(,µ)d as the probability that the next reaction in [t+, t++d] is reaction Rµ.

One can show that:

cµdt = average probability, to first order in dt, that a particular combination of Rµ reactant molecules will react accordingly in the next time interval dt. hµ = number of distinct molecular reactant combinations for reaction Rµ found to be present in V at time t. (Daniel T. Gillespie, JOURNAL OF COMPUTATIONAL PHYSICS 2, 403-434 (1976))

Example: X1 + X2 -> X3 h = X1X2

2X -> Y h = X (X-1)/2

Implementation: one has to generate (,µ) according to P(,µ) in order to update at each step the number of reactant molecules implied in reaction .

Stochastic simulations

P(,) hc exp( h1

M

c )

Daniel T Gillespie, J. Phys. Chem., 1977, 81 (25), 2340-2361

1 gene which duplicates, binomial repartitionof protein

Ages and division time distributions

Conclusion

• Build up tools in molecular biology, optic and microfluidic to measure variability in bacterial population

• Application: plasmid copy number measurement F: single copy, strictly regulated R1: partition System1) lowers PCN and 2) lowers variability ColE1: high pcn but low variability

• Plasmid loss rate in absence of partition system

• Plasmid metabolic cost: increase in division time

Perspectives• Synchronisation of bacterial population• Antibiotic concentration effect• Sorting:

• Other toxic gene to test variability

Thank you for your attention

PoubelleRéservoir 2