Indian Institute Of Technology, Bombay
Engineered Versus Natural System ENGINEERED SYSTEM Design Operation Optimization Control Engineered system: bottom-up design with known functionality of components Natural system: top down design with unknown inherent property of various motifs 11/1/2009 Team IIT Bombay, Jamboree 2009 2
Engineered Systems : Room Heater Decides to switch on/off electric supply to bring temperature to set point Temperature Controller Process Thermostat Set point for a Temperature Negative feedback Measuring Temperature SINGLE INPUT SINGLE OUTPUT (SISO) 11/1/2009 Team IIT Bombay, Jamboree 2009 3
Multiple Input Multiple Output: a motif observed in Biological System Controlled Set point Variable Process 2 Process 1 Measurement Single output is regulating the multiple upstream processes 11/1/2009 Team IIT Bombay, Jamboree 2009 4
Tryptophan in E. coli (bacteria) Ref. Venkatesh K V et al , 2004 11/1/2009 Team IIT Bombay, Jamboree 2009 5
Osmotic Stress Pathway in Yeast Ref. Parmar et al , 2009 11/1/2009 Team IIT Bombay, Jamboree 2009 6
Insulin Signaling Pathway in Mammals 7 11/1/2009 Team IIT Bombay, Jamboree 2009
Ref. Freeman, 2000 11/1/2009 Team IIT Bombay, Jamboree 2009 8
Approach Modeling – • Detailed molecular Designed and Implemented a mechanisms based synthetic genetic network model with multiple • Stochastic modeling feedbacks • Control analysis Modeling and Experiments for characterization of the network Linking Experiments protein expression • Protein expression by FACS to growth • Characterization of phenotype in the synthetic constructs 11/1/2009 Team IIT Bombay, Jamboree 2009 9
Components of Synthetic Constructs • Use of existing bio-bricks • Four promoter sites used for the constructs: pTet, pLac, pMB1 and pLacOP . • pMB1 and pLacOP : promoters for plasmid replication. • To characterize amount of LacI: LacI-CFP fusion protein. • To characterize plasmid copy number: Promoter site YFP expression. Team IIT Bombay, Jamboree 2009 10 11/1/2009
Characteristics of promoters used for Plasmid Replication On addition of IPTG Plasmid copy number does not change pMB1 On addition of IPTG Plasmid copy number increases pLacOP 11 Team IIT Bombay, Jamboree 2009 11/1/2009
LacI regulation in pTet and pLac lacI pTet pLac lacI LacI RNA/DNA Polymerase 11/1/2009 Team IIT Bombay, Jamboree 2009 12
Constructs pTet LacI +CFP pTet YFP pMB1 Plasmid Replication Plasmid Strain 1 (Open Loop, BBa_K255004) pLac LacI +CFP pTet YFP pMB1 Plasmid Replication Plasmid Strain 2 (Single Input Single Output – LacI regulation, BBa_K255003) pLacOP pTet LacI +CFP pTet YFP Plasmid Replication Plasmid Strain 3 (Single Input Single Output – Copy Number, BBa_K255002) pLacOP pLac LacI +CFP pTet YFP Plac Plasmid Replication Plasmid Strain 4 (Multiple Input Multiple Output, BBa_K255001) STOP Protein -ve Feedback Promoter 13 Team IIT Bombay, Jamboree 2009 11/1/2009
SYNTHETIC CONSTRUCTS SINGLE INPUT NO MULTIPLE INPUT CONTROL MULTIPLE OUTPUT SINGLE OUTPUT MIMO : SISO_CN : Regulation of OPEN LOOP SISO_LacI : Regulation of Plasmid Copy Regulation of Plasmid Copy (STRAIN 1) Number and LacI LacI (STRAIN 2) Number (STRAIN 4) (STRAIN 3) 11/1/2009 Team IIT Bombay, Jamboree 2009 14
Molecular Map of the Construct Replicated Plasmids LacI-IPTG complex 11/1/2009 Team IIT Bombay, Jamboree 2009 15
Modeling Methodologies • Detailed Dynamic Modeling using all known molecular interactions • Stochastic Analysis on a simplified model using Langevin approach • Frequency response analysis on the linearised model 11/1/2009 Team IIT Bombay, Jamboree 2009 16
Prediction of Steady State Expression of YFP (Plasmid Copy Number) 11/1/2009 Team IIT Bombay, Jamboree 2009 17
Control Analysis to Characterize System Behavior Controllers Block diagram for the LacI system Set-point Error LacI level + + C 1 (s) F(C s )/(s+µ+ β 1 - F’(C s ) C 1s ) k 3 C 2s /(s+µ+ β 3 ) - + Controllers C 2 (s) k 3 C ss /(s+µ+ β 3 ) Block diagram for the Linearised LacI system 11/1/2009 Team IIT Bombay, Jamboree 2009 18
Frequency Response Analysis • Higher bandwidth • Higher Phase margin • Noise Attenuation 11/1/2009 Team IIT Bombay, Jamboree 2009 19
Experimental Validation • Experiments with various IPTG concentrations were conducted. • Protein expression measured as YFP using FACS to quantify plasmid copy number. • Mean and Variance obtained from the distribution. 11/1/2009 Team IIT Bombay, Jamboree 2009 20
Experimental YFP expression (characterizing Plasmid Copy Number) Normalised YFP v/s IPTG • Open Loop and SISO_LacI: 140 No increase in YFP with 120 Normalised YFP count inducer • SISO_CN and MIMO: 100 expression increase with 80 inducer OPEN LOOP 60 SISO_pLAC 40 SISO_CN MIMO 20 Higher variance 0 0 100 200 500 -20 in open loop IPTG µM 11/1/2009 Team IIT Bombay, Jamboree 2009 21
Characterization of LacI expression • The detection of LacI-CFP fusion protein was not possible due to technical problems. • An indirect measure of LacI was obtained by measuring β -galactosidase from the lacZ of the host. • Further the growth rate of the four transformants were also enumerated. 11/1/2009 Team IIT Bombay, Jamboree 2009 22
Schematic representation of the network in presence of lactose Replicated Plasmids LacI-IPTG LacI-Lactose lacZ 11/1/2009 Team IIT Bombay, Jamboree 2009 23
Growth Response from Modeling dual feedback system: high β -gal expression with low variance 11/1/2009 Team IIT Bombay, Jamboree 2009 24
Stochastic Modeling on Growth Rate NORMALIZED β -gal SPECIFIC GROWTH EXPRESSION RATE For perturbation of the kinetic parameters around the mean value, we see MIMO has the least variance compared to open loop or a single feedback system 11/1/2009 Team IIT Bombay, Jamboree 2009 25
Experimental Results Specific Growth Rate v/s Lactose Normalized β -gal expression v/s Lactose 0.18 Specific Growth Rate (in hr -1 ) 1.2 0.16 1 0.14 β -gal/ β a-gal max OPEN LOOP 0.12 MIMO 0.8 MIMO 0.1 OPEN LOOP 0.6 0.08 0.4 0.06 0.04 0.2 0.02 0 0 0 1 2 3 4 5 6 0 1 2 3 4 5 6 -0.02 -0.2 Lactose (g/L) Lactose (g/L) Noise in protein expression propagates to growth The variance in specific growth rate is less compared to that observed in protein expression. 11/1/2009 Team IIT Bombay, Jamboree 2009 26
Agar Plate Experiments Agar Plate Experiment (without IPTG) CFU (in 10 6 /ml) 4 3 OPEN LOOP Strains were grown on agar 2 plate with different lactose 1 concentrations. 0 1 5 Lactose (g/L) Colony Forming Units in the agar plates were counted. Agar Plate Experiments (without IPTG) CFU (in 10 6 /ml) 40 MIMO 30 Variance in 20 Open Loop is 40 % and 10 MIMO is 10%. 0 1 5 Lactose (g/L) 11/1/2009 Team IIT Bombay, Jamboree 2009 27
Recapitulating… • Robustness in protein expression which leads to low variance in specific growth rate. • The noise in protein expression is filtered leading to a decrease in the variance in growth rate. This may be due to metabolism and division process. • The transformants with the synthetic network yields distinct phenotypic response. 11/1/2009 Team IIT Bombay, Jamboree 2009 28
Optimality MIMO NGR OPEN LOOP NGR Optimal production of enzyme : growth rate for MIMO. MIMO has optimized its burden for optimal Normalized Growth Rate. 11/1/2009 Team IIT Bombay, Jamboree 2009 29
IMPROVED PERFORMANCE PRECISION OPTIMALITY MULTIPLE FEEDBACK SYSTEM ROBUSTNESS FASTER TO INTRINSIC RESPONSE NOISE TIME 11/1/2009 Team IIT Bombay, Jamboree 2009 30
Acknowledgements Sponsors: Mentors: • Prof. K V Venkatesh IIT Bombay • Prof. Sharad Bhartiya • Prof. Vishwesh Kulkarni Contributors: • Mukund Thattai, NCBS, Bangalore • Dr. Manjula Reddy, CCMB, Hyderabad Phd mentors: • Pushkar Malakar, • Navneet Rai , • Vinay Bavdekar, 11/1/2009 Team IIT Bombay, Jamboree 2009 31
Thank you!! 11/1/2009 Team IIT Bombay, Jamboree 2009 32
Q&A • Bode plot analysis 11/1/2009 Team IIT Bombay, Jamboree 2009 33
Magnitude and Phase Bode plots ZERO IPTG HIGH IPTG BACK 11/1/2009 Team IIT Bombay, Jamboree 2009 34
Recommend
More recommend
